Georgia State University

ScholarWorks @ Georgia State University

Computer Information Systems Dissertations Department of Computer Information Systems

12-7-2010

What Support Does Information and

Communication Technology (ICT) Offer to
Organizational Improvisation During Crisis
Response ?
Anouck Adrot

Follow this and additional works at: http://scholarworks.gsu.edu/cis_diss

Recommended Citation
Adrot, Anouck, "What Support Does Information and Communication Technology (ICT) Offer to Organizational Improvisation
During Crisis Response ?." Dissertation, Georgia State University, 2010.
http://scholarworks.gsu.edu/cis_diss/46

This Dissertation is brought to you for free and open access by the Department of Computer Information Systems at ScholarWorks @ Georgia State
University. It has been accepted for inclusion in Computer Information Systems Dissertations by an authorized administrator of ScholarWorks @
Georgia State University. For more information, please contact scholarworks@gsu.edu.
 UNIVERSITE PARIS-DAUPHINE
EDOGEST
DRM
GFR-CREPA-MANAGEMENT&ORGANISATION
&
GEORGIA STATE UNIVERSITY
ROBINSON COLLEGE OF BUSINESS

QUEL APPORT DES TECHNOLOGIES DE L’INFORMATION ET DE LA

COMMUNICATION (TIC) A L’IMPROVISATION ORGANISATIONNELLE
DURANT LA REPONSE A LA CRISE ?

WHAT SUPPORT DOES INFORMATION AND COMMUNICATION TECHNOLOGY

(ICT) OFFER TO ORGANIZATIONAL IMPROVISATION DURING CRISIS
RESPONSE ?

THESE
Pour l’obtention du titre de
DOCTEUR EN SCIENCES DE GESTION
(Arrêté du 07 Août 2006)

Présentée et soutenue publiquement le 07 Décembre 2010 par :

ANOUCK ADROT

Co-directeurs de thèse/co-chairs : Dr. Pierre Romelaer

Dr. Daniel Robey
Rapporteurs : Dr. David Mendonça
Dr. Christophe Roux-Dufort
Suffrageants : Dr. Veda Storey
Dr. Michael Gallivan
Dr. François-Xavier de Vaujany

  1
L’université n’entend donner aucune approbation ni improbation aux opinions émises dans le
thèses : ces opinions doivent être considérées comme propres à leurs auteurs

  2
 PERMISSION TO BORROW

In presenting this dissertation as a partial fulfillment of the requirements for an advanced degree from Georgia
State University, I agree that the Library of the University shall make it available for inspection and circulation
in accordance with its regulations governing materials of this type. I agree that permission to quote from, to
copy from, or publish this dissertation may be granted by the author or, in his/her absence, the professor under
whose direction it was written or, in his absence, by the Dean of the Robinson College of Business. Such
quoting, copying, or publishing must be solely for the scholarly purposes and does not involve potential financial
gain. It is understood that any copying from or publication of this dissertation which involves potential gain will
not be allowed without written permission of the author.

Anouck Virginie Vanessa ADROT

  3
 NOTICE TO BORROWERS

All dissertations deposited in the Georgia State University Library must be used only in accordance with the
stipulations prescribed by the author in the preceding statement.

The author of this dissertation is:

ANOUCK ADROT
15, villa victorien Sardou
75016 Paris, France

The directors of this dissertation are:

Daniel Robey
CIS Deparment
9th Floor, 35 Broad Street,Atlanta, 30302, Georgia, USA

Pierre Romelaer
GFR Management & Organisation
Place du Maréchal Delattre de Tassigny, 75116 Paris Cedes, FRANCE

  4
 WHAT SUPPORT DOES ICT (Information and Communication Technology) OFFER TO
ORGANIZATIONAL IMPROVISATION DURING CRISIS RESPONSE?

BY

Anouck Virginie Vanessa ADROT

A Dissertation Submitted in Partial Fulfillment of the Requirements for the Degree

Of

Doctor of Philosophy

In the Robinson College of Business

Of

Georgia State University

GEORGIA STATE UNIVERSITY

ROBINSON COLLEGE OF BUSINESS
2010

  5
 Copyright by
Anouck Virginie Vanessa ADROT
2010

  6
 ACCEPTANCE

This dissertation was prepared under the direction of the Anouck ADROT Dissertation Committee. It has been
approved and accepted by all members of that committee, and it has been accepted in partial fulfillment of the
requirements for the degree of Doctoral of Philosophy in Business Administration in the Robinson College of
Business of Georgia State University.

H. Fenwick Huss, Dean

DISSERTATION COMMITTEE

Dr. Daniel Robey

Dr. Pierre Romelaer

Dr. Veda Storey

Dr. Michael Gallivan

Dr. David Mendonça

Dr. Christophe Roux-Dufort

Dr. François-Xavier de Vaujany

  7
 ABSTRACT

WHAT SUPPORT DOES ICT (Information and Communication Technology) OFFER TO

ORGANIZATIONAL IMPROVISATION DURING CRISIS RESPONSE?

BY

Anouck Virginie Vanessa ADROT

7th of December 2010

Committee Co-Chair: Dr. Daniel Robey & Dr. Pierre Romelaer

Major Academic Unit: Computer & Information Systems

While evidence of the exceedingly important role of technology in organizational life is commonplace,
academics have not fully captured the influence of Information and Communication Technology (ICT) on crisis
response. A substantive body of knowledge on technology and crisis response already exists and keeps
developing. Extensive research is on track to highlight how technology helps to prepare to crisis response and
develop service recovery plans. However, some aspects of crisis response remain unknown.
Among all the facets of crisis response that have been under investigation for some years, improvisation still
challenges academics as a core component of crisis response. In spite of numerous insights on improvisation as a
cognitive process and an organizational phenomenon, the question of how improvisers do interact together while
improvising remains partly unanswered. As a result, literature falls short of details on whether crisis responders
can rely on technology to interact when they have to improvise collectively. This dissertation therefore brings
into focus ICT support to organizational improvisation in crisis response in two steps: We first address this
question from a general standpoint by reviewing literature. We then propose an in depth and contextualized
analysis of the use of a restricted set of technologies – emails, faxes, the Internet, phones - during the
organizational crisis provoked by the 2003 French heat wave.
Our findings offer a nuanced view of ICT support to organizational improvisation in crisis response. Our
theoretical investigation suggests that ICTs, in a large sense, allow crisis responders to improvise collectively. It
reports ICT properties - graphical representation, modularity, calculation, many-to-many communication, data
centralization and virtuality – that promote the settling of appropriate conditions for interaction during
organizational improvisation in crisis response. In the empirical work, we provide a more integrative picture of
ICT support to organizational improvisation in crisis response by retrospectively observing crisis responders’
interactions during the 2003 French heat wave. Our empirical findings suggest that improvisation enables crisis
responders to cope with organizational emptiness that burdens crisis response. However, crisis responders’
participation in organizational improvisation depends on their communicative genres. During the 2003 French
heat wave crisis, administrative actors who had developed what we call a “dispassionate” communicative genre
in relation to their email use, barely participated in organizational improvisation. Conversely, improvisers
mainly communicated in what we call a “fervent” communicative genre. Therefore, our findings reveal that the
ICT support to organizational improvisation in crisis response is mediated by the communication practices and
strategies that groups of crisis responders develop around ICT tools.

  8
 Acknowledgements
In August 2006, I had the opportunity to initiate this dissertation project that I am about to
defend. This research has been a tremendous adventure, both from an intellectual and
personal perspective. I owe many thanks to all the people who were involved directly and
indirectly in this project.
First of all, I was honored that the committee members accepted to guide my through this
research and participate in the proposal and final defenses. Each member of this committee
contributed in their manner to the content that I will defend on the 7th of December 2010.
I especially thank my co-advisors, Pierre Romelaer at Paris-Dauphine and Daniel Robey at
GSU. Without their presence, this project would not have been possible. Their mentoring was
essential for my understanding of scientific work and the developing skills necessary to
complete a dissertation.
I am particularly grateful to François-Xavier de Vaujany and David Mendonça whose
comments had significant impacts on my work. I thank David Mendonça and Christophe
Roux-Dufort for accepting to do the difficult job of rapporteurs. I thank Michael Gallivan and
Veda Storey for their inputs during the proposal defense that were essential to the refinement
of my thinking.
I would like to express my gratitude to many other people who played a significant role in my
dissertation.
Dr. Barhadwaj and Dr. Rafaeli and the participants of the 2008 ICIS Doctoral Consortium
deserve my sincerest thanks for their insight into my work. I thank Dr. Cucchi and Dr.
Marciniak as well as the 2009 AIM Doctoral Consortium participants for their input.
In July 2006, Dr. Kalika accepted me into the dual doctoral degree program between Paris-
Dauphine and Georgia State University. I owe him many thanks for believing in my ability to
complete this dissertation. I thank Dr. Monod as well for establishing the partnership
between GSU and Paris-Dauphine.
Many thanks go to Stéphanie Dameron, Henri Isaac, Christophe Elie-dit-Cosaque and
Alexandre Renaud who discussed my work in the CREPA seminars. Their suggestions never
left my mind.
I was lucky enough to be surrounded by supporting and kind friends who agreed to revise the
content of this dissertation or provide precious inputs. An enormous thanks goes to Lionel
Garreau and Gaetan Mourmant who were wonderful companions. I had the chance to meet
many other generous persons. I experienced beautiful times in GSU and Dauphine thanks to
Chad Anderson, Claudia Ruiz, Hiro Takeda, Benoît Raymond, Mathieu Chauvet, Loréa
Hirèche, Hanane Beddi, Mantiaba Coulibaly, Antoine Harfouche and Philippe Mouricou.
Finally, I would like to thank Rachel for her tremendous help in English.

  9
I thank my family and friends for supporting me and making my personal life easier when the
dissertation required additional efforts. I thank Ngoc Tran Pham for her inspiring teaching of
meditation. I also thank Fatima for her devotion in taking care of my little one.
Finally, I am grateful to the two loves of my life, who relentlessly surrounded me with
affection during the dissertation project. Sébastien supported me in the effort of searching,
questioning, coding, reading and writing on improvisation and technology. Calixte joined us
in the midst of this effort. Since then he has provided us many opportunities for joyful
improvisation.

  10
 Index
Acknowledgements................................................................................................................... 9
Index ........................................................................................................................................ 11
Table Index ............................................................................................................................. 16
Figure Index............................................................................................................................ 17
Tableau Index ......................................................................................................................... 18
L’apport des Technologies de l’Information et de la Communication (TIC) à
l’improvisation organisationnelle durant la réponse à la crise : Un résumé en Français
de notre travail doctoral ........................................................................................................ 20
Section 1. Les objectifs et la structure de la thèse.......................................................... 21
1.I. La crise organisationnelle .................................................................................... 22
1.II. L’improvisation organisationnelle ...................................................................... 23
1.III. Les Technologies de l’Information et de la Communication ............................ 27
Section 2. Présentation des trois études qui composent la thèse.................................... 29
Section 3. Présentation des études ................................................................................. 31
3.I. Première étude: Une analyse structurée du travail académique sur l’improvisation
31
3.II. Deuxième étude: Propositions théoriques et revue de littérature sur
l’improvisation organisationnelle et les TIC................................................................ 36
3.III. Troisième étude: Les TIC, l’improvisation, le vide organisationnel : L’influence
des genres de communication sur l’improvisation de crise.......................................... 40
3.III.1. L’adéquation de la technologie à l’improvisation et les théories du fit ...... 41
3.III.2. Les résultats................................................................................................. 45
3.IV. Discussion des résultats empiriques................................................................... 50
Section 4. Les considérations finales de la thèse ........................................................... 50
4.I. Confrontation des résultats empiriques et théoriques : le modèle final de la thèse
50
4.II. Discussion des propriétés des TIC à l’aune de l’analyse empirique................... 51
4.III. Implications managériales de nos résultats........................................................ 54
Section 5. Contribution et opportunités de recherche .................................................... 55
5.I. Contributions de la recherche .............................................................................. 55
5.II. Quelques pistes de recherche .............................................................................. 55
Research questions and document overview ....................................................................... 57
Part 1. The dissertation basic features: Rationale, structure, ontology ............................ 60
Chapter 1.1 Our rationale for this research .................................................................... 61
Section 1. Needs for further research on crisis response ............................................... 62
Section 2. Some uncertainties about ICTs support to crisis response............................ 63
Section 3. Organizational improvisation: some theoretical challenges ......................... 65
Section 4. The rationale for studying improvisation in crisis response ......................... 67
Section 5. The IS researchers’ perspective on improvisation ........................................ 67
Section 6. The Dissertation topic: ICTs and organizational improvisation during crisis
68
Annex ....................................................................................................................... 73
Chapter 1.2 Four questions and three papers: Our rationale for the structure of the
dissertation.......................................................................................................................... 76
Section 1. Our rationale for the dissertation structure.................................................... 77

  11
 1.I. Three perspectives on ICT ................................................................................... 77
1.II. ICT Support in crisis response............................................................................ 79
Section 2. Interactions as unit of analysis: Our rationale............................................... 80
2.I. What is an interaction?......................................................................................... 80
2.II. The importance of interactions in crisis response and crisis improvisation ....... 83
2.III. Interaction: a convenient unit of analysis to observe ICT ................................. 84
Section 3. Research questions specifications and studies overview .............................. 84
3.I. Chapter 2.2. How to cope with polysemous concepts in organization studies? The
case of improvisation ................................................................................................... 85
3.II. Chapter 2.3 ICT, crisis and organizational improvisation: some theoretical
propositions and a literature review ............................................................................. 85
3.III. Chapter 3.3 ICT, organizational improvisation and emptiness during the 2003
French heat wave crisis response ................................................................................. 85
Chapter 1.3 Ontological and epistemological Assumptions ........................................... 88
Section 1. A pragmatist posture…. ................................................................................ 88
Section 2. …that calls for a realist ontology .................................................................. 89
Section 3. Some epistemological implications............................................................... 90
Part 2. The Theoretical assumptions and proposal for theory .......................................... 93
Chapter 2.1. The context of investigation: Presentation of the scope of enquiry and
theoretical assumptions on the concepts of crisis and crisis response........................... 94
Section 1. Crisis research as a developed but heterogeneous field ................................ 94
Section 2. Our basic assumptions on crisis .................................................................... 97
2.I. Crisis: an organizational phenomenon................................................................. 97
2.II. Social and natural Costs...................................................................................... 98
2.III. Surprise .............................................................................................................. 98
2.IV. Emotional and time pressure.............................................................................. 98
2.V. An objective and subjective existence ................................................................ 99
2.VI. Organizational crisis as a cycle........................................................................ 100
Section 3. Crisis response: some theoretical background ............................................ 102
3.I. Crisis response: a paradoxical process............................................................... 102
3.II. Improvisation in crisis response........................................................................ 103
Section 4. Crisis response in the context of natural disasters ...................................... 103
Chapter 2.2 Coping with diversity in research on improvisation................................ 106
Introduction .................................................................................................................... 106
Section 1. Theoretical background............................................................................... 108
1.I. Diversity and knowledge building ..................................................................... 108
1.II. Diversity and research on improvisation .......................................................... 110
1.III. Strategies to analyze diversity ......................................................................... 110
Section 2. Method ........................................................................................................ 112
2.I. Classification process......................................................................................... 113
2.II. Classification grid ............................................................................................. 115
2.II.1. Period.......................................................................................................... 116
2.II.2. Epistemology/paradigmatic posture ........................................................... 117
2.II.3. Reasoning about improvisation .................................................................. 117
2.II.4. Levels of analysis ....................................................................................... 118
2.II.5. Contribution to knowledge on improvisation............................................. 118
2.II.6. Research design.......................................................................................... 119
2.II.7. Focus on improvisation practices ............................................................... 119
2.II.8. Scope of enquiry......................................................................................... 119
2.II.9. Conditions of improvisation emergence..................................................... 120

  12
 2.II.10. Improvisation contexts ............................................................................. 120
2.II.11. Hierarchical status of improvisers ............................................................ 120
Section 3. Statistical treatment ..................................................................................... 121
3.I. Descriptive statistics .......................................................................................... 121
3.II. Principal Component Analysis ......................................................................... 125
3.III. Cluster analysis ................................................................................................ 128
Section 4. Interpretation of findings............................................................................. 131
4.I. Evidencing improvisation .................................................................................. 132
4.II. Focalizing.......................................................................................................... 134
4.III. Contextualizing................................................................................................ 135
4.IV. Creating knowledge ......................................................................................... 136
Section 5. Discussion of findings................................................................................. 138
Annex 1. Description of the sample ....................................................................... 142
Annex 2. Description of the requests ..................................................................... 149
Annex 3. Statistical outputs.................................................................................... 150
Annex 4. An example of replication ...................................................................... 153
Chapter 2.3 Information and Communication Technology (ICT), crisis, and
organizational improvisation: theoretical propositions and literature review........... 155
Introduction .................................................................................................................... 155
Section 1. Theoretical background on improvisation .................................................. 156
1.I. Improvisation definitions ................................................................................... 156
1.II. Some conceptual limitations ............................................................................. 157
Section 2. Organizational improvisation constituents.................................................. 159
2.I. Collective process .............................................................................................. 160
2.II. Spontaneity ....................................................................................................... 161
2.III. Deviation from established uses of resources.................................................. 161
2.IV. Interaction between actors ............................................................................... 163
Section 3. Method ........................................................................................................ 164
Section 4. Results ......................................................................................................... 168
Section 5. Discussion ................................................................................................... 171
Annex ..................................................................................................................... 174
Part 3. The empirical part ................................................................................................... 176
Chapter 3.1. Presentation of the 2003 French heat wave case ..................................... 177
Section 1. The response network actors ....................................................................... 178
1.I. Health care organizations................................................................................... 179
1.I.1. National organizations................................................................................. 180
1.I.2. Regional organizations ................................................................................ 181
1.I.3. Local organizations ..................................................................................... 182
1.II. Civil protection & internal affairs organizations .............................................. 183
1.II.1. National organizations................................................................................ 183
1.II.2. Local organizations .................................................................................... 184
1.III. “Outsider” organizations.................................................................................. 184
1.III.1. National organizations............................................................................... 184
1.III.2. Local organizations ................................................................................... 186
Section 2. The response network values ...................................................................... 186
2.I. Collaboration...................................................................................................... 186
2.II. Hierarchy........................................................................................................... 187
2.III. Organizational and informational barriers....................................................... 187
Section 3. The 2003 French heat wave as an organizational crisis .............................. 188
Section 4. Crisis response chronology ......................................................................... 191

  13
 4.I. Disparate signals (from the 4th to the 7th)........................................................... 191
4.II. The beginning of the coping (between the 8th and the 10th) ............................. 191
4.III. What do we do now? (between the 11th and the 13th)...................................... 192
4.IV. The beginning of the controversy (between the 13th and the 20th)................... 192
Chapter 3.2. The research design of the empirical investigation................................. 194
Section 1. Our research objectives ............................................................................... 195
Section 2. Method: An overview ................................................................................. 196
2.I. Our motivation for a qualitative case study ....................................................... 197
2.II. A grounded theory approach............................................................................. 197
2.II.1. Our rationale for a grounded theory approach ........................................... 197
2.II.2. A Straussian perspective ............................................................................ 198
2.II.3. Methodological implications of our following the grounded theory
approach .................................................................................................. 199
Section 3. Instrumentation ........................................................................................... 200
3.I. Retrospective observation .................................................................................. 201
3.II. Observing ICT and improvisation through interactions ................................... 203
3.III. Theoretical Sampling....................................................................................... 204
3.IV. Reality check.................................................................................................... 205
Section 4. Data sources ................................................................................................ 205
4.I. Public hearings................................................................................................... 206
4.II. Internal archives................................................................................................ 207
4.III. External archives.............................................................................................. 208
Section 5. Data handling .............................................................................................. 214
5.I. An iterative process............................................................................................ 214
5.II. Data reduction................................................................................................... 215
Annex 1. Reality check tool ................................................................................... 219
Annex 2. Document summary forms for secondary data....................................... 221
Annex 3. Dictionary themes................................................................................... 223
Chapter 3.3 ICT, communicative genres, organizational improvisation and emptiness
during the 2003 French heat wave crisis response ........................................................ 229
Introduction .................................................................................................................... 229
Section 1. Theoretical background on improvisation and crisis response ................... 230
1.I. Organizational crisis during disasters ................................................................ 230
1.II. Organizational improvisation for crisis response ............................................. 231
1.III. Technological fit to crisis improvisation ......................................................... 232
Section 2. Methodology and coding............................................................................. 235
Section 3. Findings....................................................................................................... 238
3.I. Organizational emptiness and organizational improvisation ............................. 239
3.I.1. Material resource ......................................................................................... 240
3.I.2. Immaterial resources ................................................................................... 241
3.I.3. Processes ..................................................................................................... 241
3.I.4. Instructions and orders ................................................................................ 242
3.I.5. Outputs ........................................................................................................ 242
3.I.6. Norms .......................................................................................................... 243
3.I.7. Skills ............................................................................................................ 243
3.I.8. Communicational connections between individual and groups .................. 244
3.II. Communicative genres during the 2003 French heat wave .............................. 246
3.II.1. The fervent communicative genre .............................................................. 247
3.II.2. The dispassionate communicative genre ................................................... 250
3.III. Moving from the dispassionate communicative genre.................................... 255

  14
 Section 4. Discussion of findings................................................................................. 256
Annex ..................................................................................................................... 259
Part 4. Conclusion part: Summary of findings & final discussion .................................. 264
Chapter 4.1. Dissertation summary................................................................................ 265
Section 1. The dissertation structure and its research questions .................................. 265
Section 2. Presentation of findings............................................................................... 266
2.I. Coping with diversity in research on improvisation .......................................... 266
2.II. ICT support to organizational improvisation in crisis response: a literature
review and theoretical propositions............................................................................ 267
2.III. ICT, communicative genres, organizational improvisation and emptiness during
the 2003 French heat wave crisis response ................................................................ 269
Section 3. Connections between findings .................................................................... 272
3.I. Our approach to improvisation .......................................................................... 272
3.II. ICT: Three complementary levels of analysis .................................................. 273
3.III. Final findings ................................................................................................... 274
Chapter 4.2 Conclusion.................................................................................................... 278
Section 1. Managerial contribution .............................................................................. 278
1.I. Improvisation management................................................................................ 278
1.II. ICT and communicative genres ........................................................................ 281
Section 2. Theoretical Contribution ............................................................................. 283
Section 3. Limitations and avenues for future research ............................................... 285
3.I. Improvisation ..................................................................................................... 285
3.II. Communicative genres...................................................................................... 286
3.III. Findings generalizability.................................................................................. 287
3.IV. Improvisation and dynamic capabilities .......................................................... 287
3.V. Organizational emptiness.................................................................................. 288
References ............................................................................................................................. 290

  15
 Table Index
Table 1. ICTs in crisis response ............................................................................................... 64
Table 2. Key issues that we address (or not) in the dissertation .............................................. 70
Table 3. Focus on technology and users .................................................................................. 79
Table 4. Our understanding of ICT support ............................................................................. 80
Table 5. Organizational crises and triggering events as defined by pearson and claire (1998)95
Table 6. Organizational crisis stages and corresponding issues............................................. 101
Table 7. Key characteristics of disasters, crises and crisis response...................................... 105
Table 8. Aspects of improvisation in our analysis ................................................................. 112
Table 9. Classification variables and categories .................................................................... 116
Table 10. Total variance explained ........................................................................................ 126
Table 11. Variance explained by the model - catpca on four dimensions ............................. 127
Table 12. Reliability measurement of the pca on transformed values ................................... 127
Table 13. Component loadings – catpca ................................................................................ 128
Table 14. Rotated component matrix - pca on the transformed variables.............................. 128
Table 15. Dimensions of diversity and corresponding variables ........................................... 132
Table 16. Description of the articles sample .......................................................................... 142
Table 17. Number of iterations during the catpca.................................................................. 150
Table 18. Variance accounted for each variable .................................................................... 151
Table 19. Communalities ....................................................................................................... 151
Table 20. Component transformed matrix ............................................................................. 152
Table 21. Definitions of improvisation .................................................................................. 157
Table 22. Crisis organizational improvisation constituents and their functions .................... 160
Table 23. Article extraction entry example............................................................................ 166
Table 24. Cross tabulation results .......................................................................................... 170
Table 25. Description of the sample articles .......................................................................... 174
Table 26. Major biases in our analysis................................................................................... 203
Table 27. Data sources ........................................................................................................... 206
Table 28. Organizational affiliation of the public hearings participants................................ 207
Table 29. Number of analyzed emails and faxes ................................................................... 208
Table 30. Classification of essays .......................................................................................... 210
Table 31. Classification of academic work ............................................................................ 211
Table 32. Classification of reports ......................................................................................... 212
Table 33. Clasification of newspapers article ........................................................................ 213
Table 34. The five dimensions of improvisation.................................................................... 219
Table 35. Preliminary version of the theme dictionary.......................................................... 223
Table 36. Interaction theme dictionary .................................................................................. 224
Table 37. Organizational crisis theme dictionary................................................................... 225
Table 38. Organizational improvisation theme dictionary..................................................... 226
Table 39. Organizational structures and functioning theme dictionary ................................. 226
Table 40. Individual characteristics theme dictionary ........................................................... 227
Table 41. ICT properties theme dictionary ............................................................................ 228
Table 42. Theories of fit in the is literature............................................................................ 234
Table 43. Some examples of categories ................................................................................. 236
Table 44. ICT tools, communicative genre, participation in organizational improvisation... 256
Table 45. Article projects: needs and findings....................................................................... 271
Table 46. The dissertation research questions and findings................................................... 277
Table 47. Managerial needs addressed in the dissertation ..................................................... 283
Table 48. Theoretical challenges addressed in the dissertation ............................................. 284

  16
 Figure Index
Figure 1. Les trois concepts clefs de la problématique de recherche....................................... 22
Figure 2. Structure de la thèse.................................................................................................. 28
Figure 3. Les différentes étapes de l'analyse de la littérature sur l'improvisation.................... 33
Figure 4. Les principales approches de l'improvisation en management................................. 34
Figure 5. Notre approche de l'improvisation dans la thèse ...................................................... 36
Figure 6. L'influence des propriétés des tic sur les composants de l'improvisation
organisationnelle en situation de crise ............................................................................. 39
Figure 7. Les groupes d'acteurs composant le réseau de protection civile et sanitaire en île-de-
france................................................................................................................................ 43
Figure 8. Modèle empirique ..................................................................................................... 45
Figure 9. Comparaison des genres de communication............................................................. 47
Figure 10. Impacts des genres de communication sur l'improvisation organisationnelle........ 49
Figure 11. Modèle final de la thèse .......................................................................................... 51
Figure 12. Confrontation des résultats théoriques et empiriques ............................................. 53
Figure 13. Organizational crisis, ict, organizational improvisation ......................................... 61
Figure 14. Managerial and theoretical motivations.................................................................. 62
Figure 15. Ushaidi screenshot .................................................................................................. 73
Figure 16. Instedd screenshot................................................................................................... 74
Figure 17. Ocha screenshot ...................................................................................................... 75
Figure 18. Interaction categories, bales (1950) ........................................................................ 82
Figure 19. Our rationale for the dissertation structure ............................................................. 87
Figure 20. Divergence among crisis researchers...................................................................... 97
Figure 21. Classification and analysis process....................................................................... 113
Figure 22. Improvisation context categories frequency over time......................................... 122
Figure 23. Epistemological categories frequency over time .................................................. 122
Figure 24. Reasoning categories frequencies over time......................................................... 123
Figure 25. Categories overall frequency ................................................................................ 124
Figure 26. Categories frequency over time radar................................................................... 125
Figure 27. Categories overall dendogram .............................................................................. 130
Figure 28. Eight perspectives on improvisation..................................................................... 131
Figure 29. Evidencing dendogram ......................................................................................... 134
Figure 30. Observing dendogram........................................................................................... 135
Figure 31. Contextualizing dendogram.................................................................................. 136
Figure 32. Knowledge creation dendogram ........................................................................... 138
Figure 33. Our position in relation to improvisation in the dissertation ................................ 141
Figure 34. Request screenshot................................................................................................ 149
Figure 35. Between-groups linkage dendogram .................................................................... 153
Figure 36. Highest neighborhood dendogram........................................................................ 154
Figure 37. Centroid method dendogram ................................................................................ 154
Figure 38. Death numbers (left ordinate) and celsius temperatures (right ordinate) in paris
between 6/25 and 9/15.................................................................................................... 177
Figure 39. The ile-de-france health network.......................................................................... 179
Figure 40. The 2003 heat wave crisis as an endogenous and an exogenous episode............. 191
Figure 41. Connections between topics.................................................................................. 196
Figure 42. The iterative process for theory building (from bandeira-de-mello, garreau, 2008)
........................................................................................................................................ 200
Figure 43. Data collection and analysis as an iterative process ............................................. 215
Figure 44. The conditional matrix.......................................................................................... 217

  17
Figure 45. First example of document summary forms ......................................................... 221
Figure 46. Second example of document summary forms..................................................... 222
Figure 47. The coding process steps ...................................................................................... 238
Figure 48. The overall framework.......................................................................................... 239
Figure 49. Organizational emptiness and improvisation........................................................ 245
Figure 50. Dispassionate and fervent communicative genres ................................................ 246
Figure 51. The fervent and dispassionate communicative genres and organizational
improvisation.................................................................................................................. 254
Figure 52. An AP-HP fax message ........................................................................................ 259
Figure 53. An dhos fax message ............................................................................................ 260
Figure 54. An example of email internal archive................................................................... 261
Figure 55. Another example of email archives ...................................................................... 262
Figure 56. Eight primary perspectives on improvisation ....................................................... 267
Figure 57. ICT properties and crisis organizational improvisation constituents.................... 269
Figure 58. Our approach to improvisation ............................................................................. 272
Figure 59. The final model..................................................................................................... 274
Figure 60. ICT properties and communicative genre effects on organizational improvisation
in crisis response ............................................................................................................ 276
Figure 61. A fictional example of improvisation roadmap .................................................... 280
Figure 62. Proportion of fervent and dispassionate communicative genres .......................... 287

Tableau Index
Tableau 1. Les étapes de la crise .............................................................................................. 23
Tableau 2. Les dimensions de l'improvisation ......................................................................... 25
Tableau 3. Résumé des études.................................................................................................. 30
Tableau 4. Les variables et leurs catégories............................................................................. 32
Tableau 5. Les composants de l'improvisation organisationnelle en situation de crise ........... 38
Tableau 6. Les limites de l'application des théories du fit et les motivations d'une approche
enracinée........................................................................................................................... 44

  18
« Intelligence is what you use when you don't know what to do »

Jean PIAGET,1952,
The origins of intelligence in children

  19
 L’apport des Technologies de l’Information et de la
Communication (TIC) à l’improvisation
organisationnelle durant la réponse à la crise : Un résumé
en Français de notre travail doctoral
Les organisations font face à des crises de plus en plus fréquentes et de nature très diverses:
crises financières ou politiques, catastrophes naturelles, accidents industriels et
technologiques, …
Pour répondre à ces risques, les organisations développent des systèmes de prévention et
prédéfinissent des plans d’actions. Ce mode de gestion des crises n’est pas suffisant. En effet,
suivre aveuglément les procédures préétablies de réponse aux crises n’est pas toujours
pertinent. Chaque crise est unique et ne correspond jamais à un scénario pensé a priori
(Billings et al., 1980; Quarantelli, 1988; Waugh, Streib, 2006). L’improvisation est donc une
pratique clef de la réponse à la crise. Cependant, elle reste parfois difficile à cerner pour les
managers et les chercheurs, notamment du fait de la diversité de la littérature sur ce sujet. De
nombreuses études empiriques montrent toutefois des cas d’organisations qui n’ont pas
d’autre choix que d’improviser pour répondre à une situation critique (Ciborra, 1996b;
Holden, 2005; Hutchins, 1991; Weick, 1993).
Dans ces situations, les organisations utilisent intensément les Technologies de l’Information
et de la Communication (TIC) pour rapidement mettre en place et coordonner des solutions
innovantes. En 2005, par exemple, les proches des victimes de l’ouragan Katrina ont
spontanément utilisé l’application GoogleMaps pour identifier les dommages de leurs proches
(Palen et al., 2007). Lors de la réponse au tremblement de terre en Haïti, le personnel
humanitaire a intensément utilisé les GPS et Smartphones pour répertorier les ressources
potentiellement reconvertibles dans la construction d’abris ad hoc. Malheureusement, une
large proportion de ces organisations ne parvient pas toujours à utiliser de manière efficace les
TIC dont elles disposent lorsqu’il s’agit d’improviser en situation de crise. Dans certains cas,
des redondances informationnelles peuvent créer des malentendus (Dawes et al., 2004; Jaeger
et al., 2007). Dans d’autres cas, comme celui de la réponse à la crise provoquée par la
canicule de 2003 en Île-de-France, les acteurs ne parviennent pas à improviser ensemble
lorsqu’ils interagissent par communication électronique.
On peut émettre plusieurs hypothèses pour répondre à ce questionnement. Tout d’abord il se
pourrait que l’utilisation des TIC empêche l’émergence de l’improvisation. On peut
également penser que le design des outils technologiques de communication n’est pas adapté
aux besoins des acteurs lorsque ceux-ci doivent improviser. Enfin, on peut se poser la
question de l’impact des variables individuelles et organisationnelles sur l’utilisation des TIC
durant l’improvisation de crise. Toujours est il que les managers manquent de repères sur les
tenants et les aboutissants de l’utilisation des TIC lors d’une improvisation de crise. Cette
difficulté managériale est corrélée à notre manque de connaissance sur l’improvisation et les
TIC. Cette thèse explore donc la question suivante:
  20
Quel est l’apport des Technologies de l’Information et de la Communication (TIC) à
l’improvisation organisationnelle en situation de crise ?
Nous avons décliné cette question générale en quatre questions de recherche:
I. Comment comprendre la diversité de la littérature sur l’improvisation ?
II. Dans quelle mesure les TIC, de par leurs propriétés, offrent-elles la possibilité
d’improviser lors d’une réponse à une crise ?
III. Dans quelle mesure les TIC répondent-elle aux besoins des acteurs qui participent à
l’improvisation organisationnelle en situation de crise ?
IV. Dans quelle mesure les acteurs profitent-ils de l’utilisation des TIC pour improviser
lors d’une réponse à une crise ?
Nous avons tout d’abord effectué une analyse préliminaire de la diversité de la littérature sur
l’improvisation. Cette étude nous a permis de clarifier notre positionnement conceptuel relatif
à l’improvisation. Nous avons ensuite examiné les trois autres questions par une étude
théorique puis une étude empirique. Le résumé détaille la façon dont chaque étude répond aux
questions de recherche.
Ce document est structuré de la façon suivante. Dans une première partie, nous présentons les
objectifs et la structure de la thèse. Dans un premier temps nous définissons rapidement les
termes clefs de la problématique de recherche puis les questions de recherche qui ont découlé
de cette problématique. Dans une deuxième partie, nous présentons un résumé des trois études
sur lesquelles se fonde ce travail doctoral. Enfin la dernière partie fait office de conclusion à
ce résumé. Dans un premier temps nous confrontons les résultats issus des différentes études
pour en déduire le modèle final de la thèse. Nous en présentons ensuite les implications
théoriques et managériales. Enfin nous concluons cette partie par une discussion sur les
limites de notre travail et proposons de nouvelles perspectives de recherche pour y répondre.

Section 1. Les objectifs et la structure de la thèse

La problématique de recherche fait référence à un ensemble divers de réalités

organisationnelles. C’est pourquoi nous définissons dans un premier temps les trois concepts
clefs de notre problématique de recherche, représentés par la Figure 1. Nous déduisons de
cette analyse les quatre questions de recherche qui ont guidé notre travail.

  21
 Figure 1. Les trois concepts clefs de la problématique de recherche

1.I. La crise organisationnelle

La littérature sur la crise se caractérise par un certain nombre de désaccords. D’une part, les
auteurs divergent sur les conditions d’émergence de la crise. D’autre part, certaines études
privilégient une définition objective de la crise, tandis que d’autres présentent la crise comme
une expérience avant tout subjective.
Dans notre travail, nous nous appuyons sur les approches d’Hermann (1963) et Shrivastava
(1993) pour définir ce que nous entendons par crise. Nous voyons en effet la crise comme une
série d’états que l’organisation traverse. La crise se déclenche par la rencontre de deux
phénomènes, d’une part un événement ou une situation rares et donc non anticipés par
l’organisation (Pearson, Claire, 1998), d’autre part le développement de dysfonctionnements
organisationnels (Perrow, 1999; Weick, Sutcliffe, 2001). L’exemple de l’entreprise ferroviaire
Union Pacific, Weick et Sutcliffe (2001) met évidence que la vie organisationnelle est
routinière et peut s’accompagner d’une perte de vigilance qui mène progressivement à des
dysfonctionnements organisationnels. L’incapacité de réaction de l’organisation à l’accident
ferroviaire d’Octobre 1982 en Californie résulte pour les auteurs d’un slack organisationnel
croissant sur la maintenance et le contrôle. En particulier, si l’organisation concentre son
attention sur des évènements majeurs comme des fusions ou des remaniements
organisationnels, elle est d’autant moins en mesure de gérer ses propres dysfonctionnements.
Dans un premier temps, la crise est donc est un phénomène latent, dont les signes avant-
coureurs ne sont perçus que par une minorité d’acteurs. Dans un deuxième temps, l’état de
crise est avéré par un événement déclencheur qui met évidence l’incapacité organisationnelle
à répondre à la nouvelle situation. Aussitôt se mettent en place des mécanismes individuels et
collectifs de réaction à la crise, souvent émotionnellement forts (Pauchant, Mitroff, 1992).

  22
Dans un troisième temps, l’organisation formule une réponse collective à la crise. Enfin, une
fois la crise passée, l’organisation fait également l’expérience de réactions et de processus
organisationnels post-crise, que nous ne développons pas dans ce travail.
Tableau 1 détaille les caractéristiques des quatre étapes de la crise organisationnelle mises en
évidence par Shrivastava (Shrivastava et al., 1988a) et que nous reprenons dans notre
recherche.

Tableau 1. Les étapes de la crise

Étapes de la crise Caractéristiques

Développement de Dysfonctionnements internes (Cameron et al., 1987) qui
dysfonctionnements interagissent entre eux (Perrow, 1984).
organisationnels Tendance à la rétention d’information (Beamish, 2002).
Manque de vigilance (Weick, Sutcliffe, 2001) et apparition
d’accidents (Beamish, 2002).
Réaction immédiate de Choc émotionnel proche de la mort psychologique (Pauchant,
l’organisation aux Mitroff, 1992)
évènements déclencheurs Tendance à nier la crise (Fink et al., 1971) et rigidités
organisationnelles (Staw et al., 1981 Vertinsky, 1977).
Importante incertitude (Milburn et al., 1983b 1980).
Risque d’une renversement politique (Pauchant, Mitroff, 1992).
Période de réponse Mise en place de la cellule de crise (Mitroff et al., 1996 1996;
Smart, Vertinsky, 1977).
Tendance à la centralisation (Hermann, 1963 1981) ou faible
direction (Weick, 1988).
Structures redondantes (Roberts, 1990b).
Actions désespérées sans être certain de son effet sur la crise
(Starbuck et al., 1978, cited by Fiol, Lyles, 1985).
De nouveaux processus informationnels (Osborn, 1998).
Besoins informationnels croissants (Smart, Vertinsky, 1984).
Remise en question de l’utilité des plans (Pauchant, Mitroff,
1988), l’improvisation peut alors apparaître.
Adaptation des processus organisationnels (Fink et al., 1971;
Leifer, Huber, 1977).
Période post crise Apprentissage post-crise (Fauchart, 2006; Roux-Dufort, 2007).
Incertitude de l’impact sur le long-terme de la crise (Pauchant,
Mitroff, 1992).
1.II. L’improvisation organisationnelle
l’improvisation en tant que phénomène organisationnel est difficile à observer (Crossan et al.,
2005). Pour mieux approcher le concept d’improvisation et la problématique de recherche,
nous proposons d’analyser les interactions qui composent l’improvisation organisationnelle.
Ces interactions peuvent être concrètes ou virtuelles. Nous nous référons dans cette thèse à la
définition des interactions par Bales (1950).

  23
Avant même que l’improvisation ne concentre l’attention d’une partie de la communauté de
recherche en gestion, elle avait déjà été évoquée en stratégie (Mintzberg, McHugh, 1985) et
dans le contexte des crises (Hutchins, 1991). Le numéro spécial dans Organization Science en
1998 marque l’intérêt croissant des chercheurs pour ce concept. Toutefois, il existe un nombre
important de définitions de l’improvisation. Au gré de ces définitions, cinq caractéristiques de
l’improvisation se distinguent. Nous nous sommes appuyés sur ces dernières pour identifier
l’improvisation dans notre étude empirique. L’improvisation correspond à i) une
recombinaison des ressources ii) de façon créative, iii) dans un délai court entre la décision et
l’action, permettant iv) une adaptation de l’organisation v) par l’innovation. Ces dimensions
de l’improvisation sont présentées dans le Tableau 2.

  24
 Tableau 2. Les dimensions de l'improvisation

Définitions de Dimensions de l’improvisation

l’improvisation Contradiction Innovation Créativité Recombinaison Adaptation
temporelle de ressources
“A process of X
introducing changes to
an initial model while
maintaining continuity
in the performance”
(Preston, 1987)
“Thinking and action X
emerge
simultaneously and on
the spur of the
moment” (Ciborra,
1996a)
“Convergence X X
between acting and
planning, and
reworking of pre-
composed material in
relation to
unanticipated ideas”
(Vendelø, 2009)
“Organizational X
improvisation refers to
the convergence of
conception and
execution” (Moorman,
Miner, 1998a)
“Composing and X X
performing
extemporaneously,
conception as action
unfolds” (Berliner,
1994, cited by
Bansler, Havn, 2003)
“Unexpected X X
practices, deviation
from routines,
puzzling or random
behaviors calling for
interpretation”
(Lanzara, 1999)
“Action is taken in X
spontaneous and
intuitive fashion”

  25
(Crossan, 1998)
“Action that fills the X X X
gap between routine
organizational
procedures and events
in the course of daily
work” (Suchman,
2007)
“To cope or X X
ingeniously adapt to a
set of circumstances”
(Preston, 1991, cited
by Vera, Crossan,
2005)
“Decision as action X
unfolds”. (Moorman,
Miner, 1998a)
“Devising resourceful X X
solutions to intractable
problems”. (Meyer,
1998).
“The conception of X X X
action as it unfolds by
an organization and/or
its members drawing
on available material,
cognitive, affective
and social resources”
(Cunha et al., 1999c
1999)
“Improvisation is the X X X
ability to recombine
chunks of past
experience into new
patterns of action”
(Rerup, 2001)
Comme l’illustre la variété des définitions de l’improvisation, le concept d’improvisation a
fait l’objet d’analyses et de conceptualisations variées. De plus l’improvisation a été étudiée
dans des contextes empiriques différents, allant de la réponse à la crise (Mendonça et al.,
2006; Webb, F-R., 2006), au processus de certification des chirurgiens (King, Ranft, 2001),
ou encore au fonctionnement de l’administration Israélienne (Sharkansky, Zalmanovitch,
2000). Du fait de cette diversité, nous proposons dans le premier temps de la thèse d’explorer
la question de recherche suivante :
I. Comment comprendre la diversité de la recherche sur l’improvisation ?

  26
 1.III. Les Technologies de l’Information et de la Communication
Dans cette thèse, nous développons trois niveaux d’analyse des TIC. Tout d’abord nous nous
intéressons aux TIC en tant qu’ un même objet dont les propriétés intrinsèques sont reconnues
socialement (Orlikowski, Iacono, 2001). Une analyse plus précise de la technologie consiste à
examiner les fonctions auxquelles la technologie répond dans les organisations, telle que la
coordination ou l’aide à la décision (Carver, Turoff, 2007). Nous nous intéressons alors à
l’ensemble des fonctionnalités technologiques qui permettent les interactions entre
utilisateurs, tels que la communication électronique, la messagerie instantanée, les
fonctionnalités de participations aux forums ou la transmission de documents par fax. Enfin,
nous observons les TIC au travers l’utilisation d’outils de communication électronique.
Au fil des questions de recherche, nous abordons les trois niveaux d’analyse des TIC. Notre
progression conceptuelle de la question II à la question IV suit un degré croissant de précision
d’analyse de la technologie. Elle suit également une prise en compte croissante du contexte de
l’utilisateur dans notre réflexion. En effet la question II implique un examen de la technologie
comme un objet distinct des ses utilisateurs. La question III examine l’adéquation d’une
fonctionnalité technologique à un besoin organisationnel. Enfin, la question IV analyse la
technologie au travers des pratiques. Figure 2 récapitule la structure de la thèse.

  27
 Figure 2. Structure de la thèse

  28
 Section 2. Présentation des trois études qui composent la thèse

Notre première étude répond à la question de recherche I et consiste en l’analyse de la

littérature sur l’improvisation par la mobilisation des techniques d’analyse factorielle et
typologique sur un échantillon de 105 articles. A partir de cette analyse, nous mettons en
évidence quatre aspects de la diversité de la littérature sur l’improvisation et nous proposons
une première description des approches de l’improvisation.
La deuxième étude explore d’un point de vue théorique la question de recherche II et consiste
en la proposition d’une nouvelle définition de l’improvisation organisationnelle inspirée de
l’approche des capacités dynamiques (Teece et al., 1997; Zollo, Winter, 2002). A partir de
cette nouvelle définition, nous faisons une revue de littérature aboutissant à des propositions
théoriques sur l’apport des TIC à l’improvisation organisationnelle pendant la réponse à la
crise.
Enfin la troisième étude explore les questions de recherche III et IV et consiste en une analyse
enracinée du cas de la réponse à la crise provoquée par la canicule de 2003 en Ile de France.
Un modèle intégrateur a ainsi émergé de notre analyse. Il articule les concepts de technologie,
improvisation, crise et vide organisationnel.
Dans l’ensemble de la thèse, nous adoptons un positionnement pragmatiste (Cherryholmes,
1992) qui appelle à un positionnement ontologique et épistémologique réaliste, ou depth
realist (Blaikie, 2007; Outhwaite, 2003).
Le Tableau 3 présente un résumé des besoins qui ont motivé les différentes études et les
résultats issus de notre travail doctoral.

  29

Titre en Anglais de l’étude
1. Coping with diversity in
research on improvisation
2. ICT, Crisis, and
Organizational Improvisation:
Some theoretical reviews and
a literature review
3. ICT, communicative
genres, organizational
improvisation and emptiness
during the 2003 French heat
wave crisis response
30
Besoins
1) Un moyen de comparer les études
existantes sur l’improvisation
2) Une vision claire des différentes
approches de l’improvisation
1) Manque de compréhension du rôle
des interactions dans l’improvisation
organisationnelle
2) Une analyse de l’influence des TIC
sur le déroulement de l’improvisation
organisationnelle
1) Évaluer l’adéquation des TIC à
l’improvisation organisationnelle en
situation de crise
2) Comprendre quelle était l’influence
de l’utilisation des TIC sur
l’improvisation organisationnelle
durant la réponse à la crise de la
canicule de 2003.
Résultats
La diversité de la recherche sur l’improvisation provient
de divergences entre les études sur quatre points clefs du
processus de recherche
i) Mise en évidence conceptuelle et empirique de
l’improvisation
ii) Les aspects de l’improvisation observés
iii) La contextualisation de l’improvisation
iv) La création de connaissances sur l’improvisation
1) 5 mécanismes organisationnels composent
l’improvisation organisationnelle en situation de crise
2) 6 propriétés des TIC ont un effet positif sur
l’improvisation organisationnelle en situation de crise
Tableau 3. Résumé des études
1) L’application directe des théories du fit ne permet pas
d’apporter une réponse à la question de recherche III
2) Le genre de communication dépassionné a pesé sur le
développement de l’improvisation organisationnelle
3) Au contraire, le genre de communication fervent a
facilité le développement de l’improvisation
organisationnelle.
 Section 3. Présentation des études

3.I. Première étude: Une analyse structurée du travail académique sur

l’improvisation
La première étude de notre travail doctoral consiste en une analyse systématique du traitement
du concept d’improvisation par les chercheurs en management. Notre interrogation part du
constat d’une apparente diversité parmi les travaux sur l’improvisation.
La diversité des approches d’un même concept est une préoccupation centrale dans le monde
de la recherche. Mais qu’est-ce que la diversité ? Benbasat et Weber ont mis en évidence
quatre facteurs de diversité au sein de la littérature en management (1996). Selon ces auteurs,
la diversité en sciences de gestion provient de i) de l’hétérogénéité des fondements théoriques
mobilisés, ii) de la variété des sujets ou des problèmes étudiés et iii) de l’utilisation de
différentes méthodologies.
L’approche réaliste que nous adoptons dans la thèse admet une forme de relativisme
ontologique : la diversité provient non seulement de conceptualisations différentes mais
également du fait qu’un même concept peut évoquer de multiples réalités. La manipulation
des concepts par les auteurs favorise une diversité des approches encore plus grande
(Anderson, Ortony, 1975).
La diversité d’un champ contribue sans aucun doute à sa pertinence. Selon nous, elle doit tout
de même être analysée les chercheurs. En effet, le risque de clivages entre approches d’un
même concept croît quand les chercheurs ne prennent pas en compte les différences et les
points communs entre les études qui traitent des mêmes sujet. Or il est difficile de discuter de
la diversité d’un champ sans une grille d’analyse prédéfinie. Pour cette raison, notre étude
propose une grille de classification des articles sur l’improvisation.
Pour analyser la diversité de la littérature sur l’improvisation, nous avons effectué des
analyses factorielle et typologique (Aldenderfer, Blashfield, 1984; Durrieu, Valette-Florence,
2005; Jolibert, Jourdan, 2006) sur un échantillon de 105 articles. Des requêtes sur les bases de
données EBSCO, ABI et Scholar Google nous ont permis de regrouper près de 200 articles
mentionnant le terme improvisation dans leur titre ou leur abstract. Une première lecture nous
a permis d’exclure les articles dont les abstracts mentionnaient l’improvisation mais sans
fournir une définition ou de description. Nous avons eu également recours à une
échantillonnage boule de neige pour compléter l’échantillon.
Pour élaborer cette grille d’analyse des articles de l’échantillon, nous nous sommes appuyés
sur les théories existantes sur la diversité. L’élaboration de la grille d’analyse et la définition
des modalités pour chaque catégorie ont suivi un processus itératif de discussions entre les
auteurs et de double codage. Le taux de convergence sur le double codage de la classification
est de 57%. Nous fournissons ici quelques explications des intitulés de certaines variables.
Nous avons tout d’abord segmenté la variable temps en trois périodes. La période 1 s’étend de
l’apparition des premiers articles à la date de parution du numéro spécial de la revue

  31
Organization Science sur l’improvisation. La période 2 s’étend alors de 1999 à 2003, date de
parution du numéro spécial de la revue International Journal of Management Studies sur
l’improvisation. Enfin la période 3 commence en 2004 et finit en 2010.
Nous retenons dans notre classification deux dimensions de la création de connaissance
provenant des deux questionnements suivants : la connaissance créée complète-t-elle ou
remet-elle en question les travaux antérieurs ? La connaissance créée sur l’improvisation est-
elle centrée sur le concept d’improvisation ou porte-t-elle sur des concepts connexes à
l’improvisation ?
Enfin nous distinguons deux niveaux de description de l’improvisation. Par le premier niveau
nous entendons une description du processus d’improvisation, par le second niveau nous
entendons une description de l’improvisation plus exhaustive qui prend en compte les aspects
d’innovation et d’apprentissage. Le Tableau 4 présente les variables de notre classification.

Tableau 4. Les variables et leurs catégories

Variables
Période
Le mode de réflexion sur
l’improvisation
L’approche épistémologique de
l’improvisation
L’unité d’analyse
Contribution à la connaissance
sur l’improvisation
Le design de recherche
Analyse des activités liées à
l’improvisation
Richesse de la description de
l’improvisation
Conditions d’émergence de
l’improvisation
Contexte de l’improvisation
Statut hiérarchique des individus
qui improvisent
La Figure 3 présente les différentes étapes au sein de notre analyse de la littérature sur
l’improvisation.
Catégories
Période 1(1992-1998), période 2 (1999-2004), Période
3 (2005-2010)
Déductif, Inductif, Abductif
Fonctionnaliste, Interprétativiste, Radical
Individus, Groupes, Organisation(s)
Accumulation centrée sur l’improvisation
Accumulation non centrée sur l’improvisation
Renouvellement centré
Renouvellement non centré
Étude empirique quantitative, Étude empirique
qualitative, Étude conceptuelle
Design, Utilisation
Premier niveau, Second niveau
Exogène, Endogène, Mixte
Critique, Turbulent, Normal
Bas, Moyen, Élevé, Mixte

  32
 Figure 3. Les différentes étapes de l'analyse de la littérature sur l'improvisation

Nous interprétons nos résultats statistiques de la façon suivante. Selon nous les études sur
l’improvisation divergent à quatre étapes clefs du processus de recherche. A chaque étape de
recherche, les auteurs adoptent une parmi les deux principales approches que nous avons
identifiées.
Tout d’abord les auteurs mettent en évidence l’improvisation par des raisonnements et des
méthodes distincts. Un premier groupe d’articles se caractérise par des déductions à partir d’
opérationnalisations de l’improvisation et par l’utilisation de méthodes quantitatives. Une
démarche alternative consiste à faire émerger une conceptualisation originale de
l’improvisation, sous forme d’inductions ou d’abductions à partir de cas concrets
d’improvisation. Dans ce cas, les auteurs s’appuient sur des méthodes qualitatives pour
développer leur réflexion.
  33
Ensuite, les auteurs observent des aspects différents de l’improvisation. Deux approches
principales coexistent. Certains auteurs s’intéressent à l’improvisation comme phénomène
collectif au sein du top management lors d’activités de design. D’autres auteurs analysent
l’improvisation des individus au cœur des organisations, sans distinction des niveaux
hiérarchiques.
La contextualisation de l’improvisation diffère également d’un articlée à un autre. Un premier
groupe d’articles traite de l’improvisation comme réponse à un événement inattendu et
critique. Un deuxième groupe d’articles explore l’improvisation comme mode d’organisation
favorisant le développement de nouvelles capacités.
Enfin, les ambitions des auteurs en termes de création de connaissance divergent. Un premier
groupe d’auteurs contribue à accumuler des connaissances autour d’une représentation
uniforme de l’improvisation. Au contraire, d’autres auteurs visent le développement de
nouvelles perspectives sur l’improvisation, soit par la discussion des conceptualisations
précédentes de l’organisation et de l’improvisation, soit en connectant le concept
d’improvisation à d’autres notions en management.

Figure 4. Les principales approches de l'improvisation en management

  34
Nous nous appuyons sur les résultats de cette première étude pour spécifier notre approche de
l’improvisation.
Nous adoptons tout d’abord une démarche réflexive sur l’improvisation. Nous menons dans
un premier temps une discussion sur les limites des théories sur l’improvisation
organisationnelle et formulons ensuite des propositions théoriques sur les interactions qui
composent l’improvisation organisationnelle. Dans la partie empirique de la thèse, nous
suivons les principes de la théorie enracinée et faisons émerger des données nouveau modèle
pour répondre aux questions de recherche.
Notre observation de l’improvisation se démarque des deux principales approches. En effet,
nous observons les interactions entre les individus impliqués dans l’improvisation
organisationnelle dans des activités d’utilisation des ressources.
Nous étudions l’improvisation dans un contexte de crise. L’improvisation apparaît alors
comme une réponse à une série d’évènements critiques et inattendus.
Finalement notre recherche a pour objectif de proposer de nouvelles perspectives théoriques
sur l’improvisation. La partie empirique de la thèse met en évidence les liens conceptuels
entre l’improvisation et d’autres concepts comme le vide organisationnel et les genres de
communication dépassionné et fervent.

  35
 Figure 5. Notre approche de l'improvisation dans la thèse

3.II.Deuxième étude: Propositions théoriques et revue de littérature sur

l’improvisation organisationnelle et les TIC
De nombreux auteurs soulignent l’importance des interactions lors d’une improvisation
collective (Chelariu et al., 2002). Weick (1993) et Ciborra (1996b), dans leurs analyses
respectives du désastre de Mann Gulch1, montrent qu’une mauvaise gestion des interactions
peut plonger les acteurs d’une improvisation dans le doute et faire échouer cette dernière.
Pourtant, force est de constater qu’il n’existe pas encore de description exhaustive des
interactions composant l’improvisation.
Il est en effet intéressant de noter que les définitions actuelles de l’improvisation n’intègrent
pas la notion d’interaction. Par exemple si on considère la définition de l’improvisation de
Moorman et Miner (1998a) - « decision as action unfolds » - on se rend alors compte que la
1
Durant le désastre de Mann Gulch la majorité des membres d’une équipe de pompiers moururent dans des
incendies de forêt. Le seul survivant, Dodge, a pu éviter les flammes en improvisant un feu de protection.
Malheureusement ses collègues crurent qu’il s’immolait, s’enfuirent et périrent.

  36
définition n’explicite pas comment les acteurs disposent d’un temps ou de ressources
suffisants pour se coordonner durant l’improvisation.
Pour cette raison, la première partie de notre deuxième étude a pour objectif d’expliciter les
caractéristiques de l’improvisation organisationnelle en situation de crise. Nous proposons
que l’improvisation organisationnelle est le fruit d’une déviation collective et spontanée par
rapport à l’utilisation habituelle des ressources. De plus, les interactions entre les acteurs
permettent aux groupes d’ajuster leurs actions durant l’improvisation. Trois mécanismes ont
été identifiés dans la littérature, comme partie intégrante du maintien de la cohérence de
l’action collective : le boundary spanning, le leadership par l’expertise, et le raisonnement
dialogique.
Tout d’abord, le boundary spanning (Kapucu, 2006; Tushman, Scanlan, 1981) ou interaction
de traduction permet la coordination entre des groupes d’acteurs divers (Wenger, 1999). Cette
interaction se caractérise par des processus de traduction permettant la mise en place d’un
référentiel commun d’actions (Kapucu, 2006). La mise en place de ce référentiel permet la
mise en place d’un leadership fondé sur l’expertise (McKinney, 2008; Weick, Sutcliffe,
2001). Le leadership par expertise caractérise l’écoute et l’alignement des actions d’un groupe
sur l’interprétation d’une personne experte sur une problématique précise (Weick, Sutcliffe,
2001). L’improvisation résulte également d’interactions dites de raisonnement dialogique
entre acteurs (Faraj, Xiao, 2006). Ces trois types d’interactions sont structurés par des
structures minimales de connaissance (Kamoche, Cunha, 2001) qui consistent en un ensemble
de règles simples. Le Tableau 5 donne davantage d’éléments sur les composants de
l’improvisation organisationnelle de crise.

  37
 Tableau 5. Les composants de l'improvisation organisationnelle en situation de crise

Composant de l’improvisation Spécificités

organisationnelle
Spontanéité La capacité des individus ou des groupes à formuler une
réponse à un événement dans un bref délai (George,
1997).
Déviation par rapport à Fait référence à la notion de bricolage – make do with
l’utilisation habituelle des resources at hand (Baker et al., 2003), présenté comme
ressources un composant de l’improvisation (Kamoche et al., 2003).
Interactions de traduction La capacité de certains individus ou objets à collecter,
interpréter et sélectionner les informations partagées par
plusieurs groupes d’individus (Kapucu, 2006).
Structures minimales Règles élémentaires de fonctionnement d’un groupe de
travail (Kamoche, Cunha, 2001).
Leadership fondé sur l’expertise Leadership émergent des personnes expérimentées ou
légitimes aux yeux de plusieurs collaborateurs (Weick,
Sutcliffe, 2001).
Notre deuxième étude consiste, tout comme la première, en une analyse de la littérature. Le
point de départ de notre analyse a été la recherche d’articles sur plusieurs bases de données
dont EBSCO, ABI et Web of Science. Plusieurs recoupements entre les bases de données
nous ont permis de récolter environ 150 articles, issus de la requête suivant : Technology +
response + crisis dans l’abstract. Nous ne mentionnons pas l’improvisation parmi les termes
clés de recherche. En effet, dans certains articles, l’improvisation est décrite sans être
explicitement mentionnée dans l’abstract.
Nous avons ensuite retenu les articles qui mentionnaient les éléments clés de l’improvisation
organisationnelle récapitulés dans le Tableau 5, ce qui a réduit l’échantillon à 12 articles. Pour
chaque article classé, nous avons analysé comment les auteurs décrivaient l’apport des TIC à
chaque composant de l’improvisation organisationnelle que nous présentons dans les
paragraphes précédents.
Notre analyse nous a permis de faire émerger 6 caractéristiques de la technologie que nous
interprétons comme des sources potentielles d’aide à l’improvisation organisationnelle en
situation de crise. En effet, la combinaison de ces caractéristiques permet l’expression des
composants de l’improvisation organisationnelle.
Dans ce paragraphe nous présentons les propriétés des TIC que nous avons identifiées dans ce
travail. La Figure 6 représente chaque propriété et son impact sur l’improvisation
organisationnelle en situation de crise.

  38
 Figure 6. L'influence des propriétés des TIC sur les composants de l'improvisation
organisationnelle en situation de crise

La représentation graphique, qui fait référence à une cartographie conceptuelle ou

géographique, permet l’uniformisation du format des données dans un référentiel unique
(Comfort, 1993). La technologie joue dans ce cas le rôle d’agent d’interface en forçant
l’harmonisation des informations échangées sur un même format. De plus, la représentation
graphique, par exemple sous forme de cartes, permet de concentrer diverses informations sur
un même support et d’utiliser des couleurs et des formes pour favoriser la génération
spontanée de nouvelles idées (Dantas, Seville, 2006)
La modularité des applications permet la mise en place rapide de canaux de communication
pendant la crise. Elle facilite la mise en forme de pratiques innovantes.
La puissance de calcul, la communication multilatérale et le caractère virtuel des informations
facilitent la déviation de l’usage des ressources. Par exemple, les systèmes experts comme le
Control & Command System, permettent de repérer quand le cours de l’action collective
dévie des procédures établies, ce qui facilite l’émergence de nouvelles idées pour répondre à
ces différences (Ntuen et al., 2006). Le raisonnement dialogique et la déviation par rapport
aux schémas préétablis d’action sont alors facilités.
La centralisation des données favorise la création d’une narration indispensable aux
interactions d’interface, de raisonnement dialogique et au leadership par expertise. Au travers

  39
d’une même interface, les acteurs peuvent élaborer une unique narration qui remplit la
fonction de traduction et facilite l’émergence de règles communes (Majchrzak et al., 2007). Il
est techniquement aisé de retrouver l’historique des échanges entre les acteurs ou ses propres
échanges avec les collègues. L’accès aux interactions précédentes permet de maintenir un
référentiel commun d’actions. Ce résultat confirme les conclusions des travaux précédents,
qui mettent en évidence l’importance des logs dans les échanges entre les acteurs d’une crise
(Turoff et al., 2004)
Enfin les interactions sont facilitées par le caractère virtuel des données. Les interactions de
traduction, de raisonnement dialogique et de leadership par expertise ne pourraient pas se
développer sans l’échange immédiat des informations (Comfort, 1993; Quarantelli, 1997;
Yuan, Detlor, 2005). En effet, l’échange rapide des données permet aux acteurs de la réponse
à la crise d’avoir accès à une information mise à jour presque en temps réel. Ainsi ils peuvent
concentrer leurs efforts sur la résolution des problèmes présents. Les échanges rapides
facilitent la transmission rapide des règles de coordination à tous. Ainsi les structures
minimales nécessaires à l’improvisation sont rapidement mises en place grâce aux TIC.
Sur la base de notre compréhension théorique de l’improvisation organisationnelle, nous
formulons des liens de causalité entre des propriétés des TIC et les composants de
l’improvisation organisationnelle. Nos résultats ne s’appliquent pas uniquement à un cas
particulier de réponse à une crise mais à l’ensemble des cas possibles d’improvisation en
situation de crise. L’étude met sur un même plan des approches très variées de la technologie,
qu’elles soient purement conceptuelles ou fondées sur l’analyse d’une réalité empirique,
qu’elles soient normatives ou purement descriptives.
Ce travail présente toutefois des limites qui appellent un examen approfondi des pratiques
liées aux TIC dans le cadre de l’improvisation organisationnelle en situation de crise. Nos
résultats ne décrivent qu’un apport potentiel de la technologie, qui élude l’effet de la
contingence de l’utilisateur ou de la culture organisationnelle sur les liens entre les TIC et
l’improvisation organisationnelle.
3.III. Troisième étude: Les TIC, l’improvisation, le vide organisationnel :
L’influence des genres de communication sur l’improvisation de crise.
Ce chapitre consiste en une étude détaillée des interactions qui ont participé à l’improvisation
organisationnelle durant la réponse à la canicule de 2003. En Août 2003, une vague de
chaleur s’est étendue sur l’Europe et a provoqué la mort de 14802 personnes en France. Ainsi,
le taux de mortalité en Île-de-France a doublé entre le 4 et le 20 Août 2003. Les personnes
âgées et les nourrissons ont été les plus touchés. Durant le mois d’août 2003, le réseau formé
par les organisations chargées de la protection civile et sanitaire d’Île-de-France a fait face à
une surmortalité et s’est retrouvé dans l’impossibilité de mener à bien sa mission : protéger et
sauver les vies de citoyens de la région parisienne. Les hôpitaux ont alors cruellement
manqué de ressources, les médecins assistant impuissants à l’augmentation progressive du
nombre de morts. Enfin, les sapeurs-pompiers et les services de police ont été dans
l’incapacité d’empêcher les nombreux malaises sur la voie publique ou dans les domiciles des

  40
personnes âgées. Des défaillances organisationnelles ont été mises en avant pour expliquer ce
nombre élevé de décès (Lalande et al., 2003), comme le manque de communication entre les
organismes publics de contrôle et les groupes opérationnels (services hospitaliers, médecins,
personnel des maisons de retraites, collectivités territoriales). Pour autant, certains groupes
qui n’ont pas disposé de l’information nécessaire ont dû improviser pour répondre aux
accidents et malaises croissants ainsi qu’à l’afflux de malades dans les services d’urgence.
Aussi, notre étude se concentre sur ce que nous appelons le réseau de réponse, constitué de
23 types d’organisations, tous impliqués dans la réponse à la crise provoquée par la canicule
Cette étude a pour objectif de répondre à deux des quatre questions de recherche qui
structurent notre travail doctoral, soit les questions suivantes :
III. Dans quelle mesure les TIC répond-elle aux besoins des acteurs qui participent à
l’improvisation organisationnelle en situation de crise ?
IV. Dans quelle mesure les acteurs profitent-ils de l’utilisation des TIC pour improviser lors
d’une réponse à une crise ?
3.III.1. L’adéquation de la technologie à l’improvisation et les théories du fit
Dans un premier temps, nous proposons une réflexion sur la question de recherche III qui
porte sur l’adéquation entre la technologie et les besoins des acteurs qui doivent improviser
ensemble lors de la réponse à une crise. Dans la littérature anglo-saxonne, la question de
l’adéquation entre une fonction technologique et l’utilisation d’une technologie a mobilisé un
nombre conséquent de chercheurs. Une partie de cette réflexion s’inscrit dans la continuité
des travaux sur la contingence des organisations, telle que définie par Drazin et Van de Ven
(1985). Ces théories partent de l’idée suivante : il faut un minimum de cohérence entre les
structures d’une organisation et son contexte pour atteindre un certain niveau de performance
(Drazin, Van de Ven, 1985). Plus précisément, de nombreuses théories qui relèvent de
l’approche interactionnelle du fit ont examiné l’effet du degré d’adéquation entre les
ressources technologiques et le contexte d’utilisation sur la performance de l’organisation
(Drazin, Van de Ven, 1985; Zigurs, Buckland, 1998). Ces théories prennent en compte
différents aspects du contexte de l’utilisation d’une technologie, tels que les tâches de
l’utilisateur, les besoins en termes d’interactions, la stabilité environnementale, etc.
Depuis quelques années, la recherche en gestion des crises pose elle aussi la question de
l’adéquation entre les technologies utilisées et les spécificités d’une réponse à la crise
(Mendonça, 2007; Mendonça et al., 2003). Les avis divergent fortement sur le sujet. Pour
certains auteurs, les TIC constituent un apport fondamental à la résilience. Pour d’autres la
technologie ne répond pas aux besoins les plus cruciaux et risque même de peser sur
l’improvisation en situation de crise (Dawes et al., 2004; Hiltz, Turoff, 1985; Quarantelli,
1997). Les théories du fit permettent à première vue de départager les avis et de résoudre
l’incertitude concernant l’apport d’un ensemble des TIC à l’improvisation en situation de
crise. En effet elles offrent une grille d’analyse englobant l’ensemble des variables
organisationnelles et environnementales susceptibles d’interagir avec les caractéristiques
technologiques. De cette façon elles intègrent la contingence de l’utilisateur à la réflexion en

  41
comparant précisément les besoins des utilisateurs avec les fonctionnalités d’une ou plusieurs
technologies. Pour mettre en évidence les construits susceptibles d’être intégrés à notre
réflexion sur l’adéquation entre la technologie et l’improvisation et l’improvisation de crise,
nous nous sommes inspirés de la récente revue de littérature de Zigurs et Khazanchi (2008).
Nous déduisons de cette revue de littérature les construits à prendre en compte dans l’analyse
de l’adéquation entre la technologie et l’improvisation en situation de crise.
Nous déduisons de cette analyse la difficulté d’appliquer les théories du fit au cas particulier
de l’improvisation en situation de crise. Tout d’abord, l’improvisation consiste en une
déviation spontanée et innovante des schémas d’action préétablis (Crossan, 1998; Rerup,
2001). Il est donc impossible de prédéterminer en quoi l’improvisation va consister ni les
individus qu’elle va impliquer dans l’organisation. Certes, l’improvisation peut impliquer une
complexité et une équivocité accrues. A cause de la spontanéité des acteurs et de la quasi-
absence de délai entre la décision et l’action, les acteurs ont besoin de davantage de mises à
jour régulières et détaillées sur les actions des uns et des autres. Sans aucun doute, la gestion
des interventions impromptues de bénévoles a complexifié la gestion de la sécurité et a
impliqué une redéfinition continue des flux de tâches lors du désastre du 11 septembre (Jaeger
et al., 2007). Même si la complexité et l’équivocité découlent certainement de
l’improvisation, cette dernière revêt d’autres caractéristiques qui ne sont pas prises en compte
par les théories du fit. De plus, la réponse aux crises organisationnelles causées par des
désastres se caractérise par une diversité d’organisations, de tâches et d’individus. La Figure 7
qui présente les acteurs de la réponse à la canicule de 2003 donne une idée de la diversité des
cultures organisationnelles en jeu. Les crises causées par les désastres comme les attentats du
11 septembre ou l’ouragan Katrina impliquent parfois plus de 50 organisations différentes,
incluant parmi elles des entreprise, des administrations et parfois des associations (Eisenberg,
2006). Au sein d’une même interaction, il n’est pas improbable que la technologie ne soit pas
appropriée à l’un des deux acteurs.
Tout ceci implique la nécessité d’une observation des pratiques des utilisateurs, qui est l’objet
de l’analyse de notre étude de cas. La Figure 7 présente les différents groupes d’acteurs
composant le réseau de protection civile et sanitaire d’Île-de-France : elle présente les
organisations chargées de la protection civile, les organisations chargées de la protection
sanitaire et les « outsiders », organisations qui ont développé des liens forts de coopérations
au sein du réseau de réponse durant la canicule. Les flèches en gras mettent en évidence des
liens de coopération intense entre certaines organisations. Chaque membre du réseau possède
une culture et une appréhension de l’information spécifique. Leur utilisation des systèmes
d’information et des TIC varie également : alors que les acteurs administratifs et
institutionnels privilégient l’email pour communiquer entre eux, les opérationnels
interagissent plus volontiers par téléphone. Enfin les échanges entre les sphères
opérationnelles et administratives se font relativement souvent par fax.

  42
 Figure 7. Les groupes d'acteurs composant le réseau de protection civile et sanitaire en
Île-de-France

Notre analyse a porté sur les interactions qui ont eu lieu au sein et entre les organisations qui
composent le réseau de réponse. Les limites d’une application directe des théories du fit nous
ont conduit à suivre l’approche de la théorie enracinée pour analyser ces interactions (Locke,
2001), tel que décrit par le Tableau 6.

  43
 Tableau 6. Les limites de l'application des théories du fit et les motivations d'une
approche enracinée

Limites d’une application des théories Raisons du recours à la théorie enracinée

du fit
L’absence de définition de certains des Les théories existantes sont incomplètes pour
construits empêche l’application des expliquer les phénomènes. Un enrichissement du
théories du fit à l’improvisation de crise. corpus théorique passe par une observation des
On doit alors observer l’improvisation pratiques
sur le terrain.
Les théories du fit ne prennent pas en Les connexions théoriques entre les concepts
compte les interdépendances entre la sont inexistantes ou ambigües.
technologie et l’improvisation.
Il est nécessaire de prendre en compte les Un enrichissement du corpus théorique passe par
pratiques des utilisateurs pour enrichir la une observation des pratiques ou des acteurs
réflexion sur l’adéquation d’une
technologie.
Le choix du cas à étudier s’est fait selon la technique de l’échantillonnage théorique (Mason,
2002). Dans cette perspective, nous avons sélectionné le cas d’une crise organisationnelle
impliquant une improvisation fréquente des individus. Une première comparaison entre les
données et nos hypothèses de départ sur la réponse à la crise nous a permis de reconnaître la
pression émotionnelle, la contrainte temporelle, le manque de ressource et la remise en
question des valeurs fondamentales (Hermann, 1963; Pearson, Claire, 1998). Ensuite nous
avons utilisé la technique du reality check (Harris, 2001) pour nous assurer que le cas
constituait bien un exemple d’improvisation collective. Nous nous sommes alors référés aux
cinq dimensions de l’improvisation évoquées par les différentes définitions que nous avons
trouvées dans la littérature (pour plus de détail se référer au Tableau 2). Nous avons
systématiquement comparé ces dimensions avec les différents épisodes de la réponse à la
canicule de 2003 par le personnel administratif et hospitalier du réseau. Nous avons pu alors
constater que non seulement les individus improvisaient mais que cette improvisation était
organisée à l’échelle de plusieurs groupes d’acteurs. Enfin, lors du codage des données, nous
avons suivi le principe de l’échantillonnage théorique tel que décrit par la méthode de la
théorie enracinée (Strauss, Corbin, 2008). Pour cela, nous avons élaboré des dictionnaires des
thèmes en vue de la réduction des données pour le codage.
Le cas de la réponse à la canicule de 2003 est un exemple d’improvisation organisationnelle
limitée : si certains groupes d’acteurs ont improvisé ensemble, d’autres n’ont pas participé à
l’improvisation, empêchant son développement à l’échelle du réseau. Les groupes d’acteurs
opérationnels représentés en bas de l’arborescence de la Figure 7 ont improvisé intensément
et ont utilisé les téléphones pour se coordonner. Les groupes d’acteurs administratifs
représentés en haut de l’arborescence sont, pour la plupart, restés hors de l’improvisation.
Enfin, certains acteurs administratifs ont improvisé leur utilisation du fax pour transmettre les
informations aux opérationnels.

  44
La théorie enracinée a pour objectif de faire émerger des catégories conceptuelles de la réalité
empirique, puis de les organiser.
Pour mener notre analyse, nous avons mobilisé quatre sources de données:
les entretiens publics menés lors des commissions d’enquêtes et des commissions
d’informations des chambres parlementaires,
les emails, fax et communications internes qui ont été transmises par les acteurs durant la
réponse à la crise,
les témoignages des différentes expériences de la crise, via des ouvrages ou les archives des
quotidiens,
les données secondaires provenant de travaux préliminaires sur la réponse à la canicule de
2003, sous forme de rapports nationaux et internationaux ou de travaux académiques.
3.III.2. Les résultats
Nous commençons par présenter le modèle global qui a émergé de l’analyse des données
empiriques. Nous détaillons ensuite chaque composant du modèle pour préciser notre propos.
Le modèle empirique, représenté dans la Figure 8, a émergé de l’analyse enracinée et a pour
vocation une explication holistique des rapports entre l’improvisation, le vide organisationnel
et les TIC. Nous démontrons en effet que l’improvisation a permis de combler le vide
organisationnel qui a pesé sur la réponse à la crise provoquée par la canicule de 2003.
Toutefois, la participation des individus à l’improvisation organisationnelle a été influencée
par les genres de communication qu’ils avaient développés. Nos résultats nous permettent de
répondre tout d’abord à la question de recherche IV : Dans quelle mesure les acteurs
profitent-ils de l’utilisation de la technologie pour improviser lors d’une réponse à une
crise ? Nos résultats permettent également d’apporter quelques éléments de réponse à la
question de recherche III.

Figure 8. Modèle empirique

  45
 3.III.2.1. Le vide organisationnel et l’improvisation organisationnelle
comme réponse
Notre analyse de la canicule de 2003 suggère que les organisations peuvent faire l’expérience
d’un vide organisationnel lors de la réponse à la crise. Dans ce travail, nous avons fait
émerger le concept de vide organisationnel. Le vide organisationnel caractérise une situation
dans laquelle des ressources et/ou des moyens ne peuvent pas être intégrés à un ensemble
d’actions, empêchant leur déroulement. Les acteurs font l’expérience du vide au travers du
manque de ressources (matérielles ou immatérielles) telles que définies par Barney (1991) et
le manque de moyens, tels que définis par Mintzberg (1979). Le concept de vide
organisationnel a émergé de l’analyse enracinée, mais de nombreux auteurs y font référence
plus ou moins explicitement dans leurs travaux. Alter, par exemple, met en évidence l’effort
de réflexivité font preuve les acteurs pour faire face au manque de définition des procédures
(2000). Auparavant Crozier et Friedberg ont montré comment les acteurs intègrent les zones
d’incertitude dans leur travail (1977). Weick évoque la capacité des organisations à
développer une nouvelle organisation ad hoc à partir de situations chaotiques (2001c). De
manière plus explicite, la théorie des vides structurels met en évidence l’émergence des
comportements innovants à partir de manques informationnels (Burt, 1992, 2004).
Le vide organisationnel appelle les acteurs à faire preuve d’ingéniosité et d’esprit d’initiative
pour mener à bien leur mission. Ils improvisent alors ensemble, en se coordonnant pour
développer des nouvelles pratiques collectives.
Parmi les données empiriques, nous avons identifié les composants de l’improvisation
organisationnelle que nous proposons dans le volet théorique de la thèse. La réponse à la crise
de la canicule de 2003 illustre en effet de nombreux cas d’improvisation collective. Certaines
de ces improvisations ont été spontanément orchestrées par des individus chargés de traduire
les besoins des uns aux autres et réciproquement. L’exemple d’un cadre infirmier qui s’est
spontanément chargé d’appeler les différents services pour réguler les flux de patients entre
les hôpitaux avoisinants illustre l’importance des interactions de traduction. Dans d’autres
cas, les acteurs ont spontanément mis en place une ou deux règles de conduite pour orchestrer
l’utilisation déviante des ressources. De cette façon les acteurs ont pu trouver les moyens de
mener à bien leurs tâches, malgré le manque de ressources, comme des lits, ou de moyens,
comme des processus ou des règles de gestion appropriés. Par exemple, les groupes
d’intervention d’urgence sanitaire ne savaient pas comment identifier et soulager
l’hyperthermie au début de la crise. En réponse à cela, un médecin chef de la Brigade des
Sapeurs Pompiers de Paris (BSPP) a proposé d’introduire la température des victimes comme
code d’intervention pour la BSPP, les SMUR et SAMU. Tout en utilisant une seule donnée –
la température, les acteurs ont pu facilement distinguer les interventions directement dues à la
chaleur et se coordonner pour le transfert des victimes dans les services d’urgence.
L’improvisation organisationnelle apparaît donc comme une réponse possible à l’expérience
d’un vide organisationnel. Nous allons maintenant décrire comment l’improvisation
organisationnelle a été influencée par les genres de communication existant au sein du réseau
de réponse à la canicule de 2003.
  46
 3.III.2.2. Les genres de communication
Durant le processus d’intégration décrit précédemment, des allers et retours systématiques
entre la théorie et les données ont alimenté un rapprochement des concepts avec les théories
existantes. C’est ainsi que nous avons identifié deux genres de documentation et leurs effets
sur l’improvisation organisationnelle.
Les genres de communication ont été identifiés par Yates et Orlikowski (1992). Un genre de
communication correspond à un ensemble de pratiques et de stratégies rhétoriques
progressivement développées et reconnues au sein d’une communauté d’interlocuteurs
(Orlikowski, Yates, 1994; Yates, Orlikowski, 1992). Un genre de communication résulte
d’une structuration de pratiques de communications et de l’appropriation d’un moyen de
communication. En effet, un genre de communication peut être associé à un ou plusieurs
supports, comme les rapports internes ou mémos. Il peut également être associé à un média de
communication, comme l’email (Ducheneaut, 2002; Zucchermaglio, Talamo, 2003). Le genre
de communication se caractérise par un ou plusieurs objectifs et par des éléments de forme,
allant de la structure des messages au langage adopté par les interlocuteurs (Zucchermaglio,
Talamo, 2003).
Au sein du réseau formé par les organisations impliquées dans la réponse à la crise de la
canicule de 2003, nous identifions deux principaux genres de communication. Nous les
présentons dans les paragraphes suivants, ainsi que leurs effets sur l’improvisation
organisationnelle durant la réponse à la crise. La Figure 9 présente les caractéristiques des
genres de communication dépassionnés et fervents.

Figure 9. Comparaison des genres de communication

  47
Un premier genre de communication, que nous appelons le genre fervent, prédomine parmi
les groupes d’acteurs opérationnels qui interagissent en face à face ou par téléphone durant la
réponse à la crise. Il se caractérise par une stratégie de discussion ou de persuasion, ainsi que
l’expression des émotions et opinions personnelles. Dans la forme, les objets, comme les lits
ou les brancards, sont couramment mentionnés, ce qui enrichit les discussions et favorise
l’émergence de nouvelles idées. Les émotions, telles que la colère ou la peur, ou encore les
sensations comme la soif sont exprimées sans détour et le plus souvent écoutées avec
empathie. En s’exprimant librement sur leurs émotions, les acteurs émettent leurs avis et les
idées spontanément. Ce genre de communication coïncide avec des valeurs professionnelles
orientées vers l’action, propres aux professionnels de la santé.
Le deuxième genre de communication que nous qualifions de dépassionné est plus répandu
parmi les membres du réseau en charge de tâches administratives qui échangent des emails ou
des fax. Ce genre répond à une stratégie de protection des interlocuteurs contre le risque
d’être mis à défaut par la hiérarchie ou les collègues, qui est perceptible dans le milieu
médical (Morgan, Krone, 2001). L’objectif pour les acteurs est de mettre en évidence leur
professionnalisme, notamment par leur réactivité, ainsi que leur volonté de partager les
informations. Dans la forme, les interactions par email et fax se caractérisent par une
neutralité du ton, une faible dispersion des acteurs qui restent concentrés sur le sujet de
départ. La donnée étant perçue comme une ressource objective et donc légitimant le
raisonnement des acteurs, il s’ensuit une chasse à la donnée. Ceci limite les discussions sur
l’improvisation en cours et les objets impliqués dans l’improvisation. Les échanges sont
consensuels et impliquent les personnes en fonction de leur statut dans l’organisation, ce qui
limite l’implication de personnes expérimentées mais statutairement inférieures ou
marginales, aux dépens l’émergence du leadership par expertise. Enfin, l’improvisation étant
peu mentionnée et le vocabulaire avant tout procédural, il est difficile de développer un
référentiel commun avec les acteurs qui ont improvisé durant la réponse à la crise. La Figure
10 résume nos résultats et récapitule l’influence des genres de communication sur
l’improvisation organisationnelle durant la réponse à la crise provoquée par la canicule de
2003. Les flèches (+) représentent les effets positifs tandis que les flèches (-) représentent les
effets négatifs.

  48
 Figure 10. Impacts des genres de communication sur l'improvisation organisationnelle

  49
Ainsi, les interlocuteurs administratifs qui ont avant tout échangé par email ont éprouvé des
difficultés à participer à l’improvisation en cours durant la réponse à la crise. Si
l’improvisation est pratiquement restée étrangère aux utilisateurs des emails, il est intéressant
de noter que l’utilisation du fax a permis une évolution du genre de communication
dépassionné. En effet, quelques exemples de fax annotés à la main, exprimant une émotion de
l’interlocuteur, montrent que le genre de communication dépassionné a évolué durant la crise
en ce qui concerne le fax.
3.IV. Discussion des résultats empiriques
Le modèle que nous proposons met en avant les limites de l’utilisation de l’email lors de la
réponse à la crise provoquée par la canicule de 2003 en Île-de-France. Toutefois, il est
important de noter que l’objet de notre propos n’est pas de critiquer la technologie, mais
plutôt de mettre en avant que dans certaines organisations, les habitudes développées autour
de l’email peuvent freiner l’improvisation organisationnelle. En effet, de nombreux exemples
d’utilisations inattendues d’emails, comme lors des attaques du 11 septembre, prouvent que
les TIC ne sont pas un frein systématique à l’improvisation (Jefferson, 2006b).
L’utilisation des emails a également permis de développer une narration par le suivi des
messages de par la capacité de stocker, citer ou transmettre les propos des interlocuteurs sans
effort. Notre analyse suggère que les acteurs ont intensément utilisé cette fonctionnalité de
l’email, ce qui a permis de faire remonter certaines informations et d’assurer une continuité
des thèmes abordés entre les interlocuteurs. Cette analyse confirme les résultats théoriques de
la deuxième étude, selon lesquels la centralisation des données et la virtualité de l’information
propres aux TIC permettent une narration entre les acteurs qui doivent improviser en situation
de crise (pour plus d’éléments sur la centralisation des données et la virtualité de
l’information, se référer au chapitre 2.3). Pour autant, notre analyse suggère que les
utilisateurs d’emails et de fax ne sont pas parvenus à participer à l’improvisation
organisationnelle, du fait de leurs genres de communication.

Section 4. Les considérations finales de la thèse

4.I. Confrontation des résultats empiriques et théoriques : le modèle final de la

thèse
Cette thèse a pour objectif d’expliciter l’apport des TIC à l’improvisation organisationnelle en
situation de crise. Une première exploration théorique nous a permis de mettre en évidence
que certaines propriétés des TIC facilitent l’improvisation organisationnelle. Étant collective,
l’improvisation suppose une coordination particulière, notamment grâce au raisonnement
dialogique entre experts, aux interactions de traduction et aux structures minimales de
connaissance.
Certes, la technologie facilite certaines interactions spécifiques à l’improvisation
organisationnelle. Toutefois, les groupes d’acteurs développent également des genres de
communication autour des média qui peuvent influencer l’effet des TIC sur l’improvisation
organisationnelle. La faculté de certains groupes d’acteurs à improviser dépend également de
  50
leur capacité à faire évoluer leurs genres de communication pendant la crise, ou du moins à
rendre leurs genres de communication compatibles avec ceux des autres participants de la
réponse à la crise. La Figure 11 décrit le modèle final de la thèse qui récapitule les concepts
abordés et sert de point de départ à un affinement des propositions théoriques.

Figure 11. Modèle final de la thèse

4.II. Discussion des propriétés des TIC à l’aune de l’analyse empirique

Nous proposons ici une discussion des résultats théoriques à partir notre analyse empirique.
Notre réflexion met en évidence que les propriétés des TIC ne sont pas seulement des traits
essentiels de l’artefact technologique mais sont également perçues et interprétées par les
utilisateurs. Notre propos se concentre ici sur la technologie de l’email.
Le cas de la canicule de 2003 n’illustre pas d’utilisation de représentation graphique dans les
emails ou les fax. Les données, tout de même, mettent en évidence que certains acteurs se
sont mis à improviser par un contact proche avec la réalité de la crises, autrement dit après
avoir vu, pour de vrai, les personnes étendues sur les brancards et en manque de soin. On
pourrait alors faire l’hypothèse que les images faciliteraient les échanges sur les émotions, les
opinions ou les objets en jeu lors de l’improvisation organisationnelle. Ces propositions
nécessitent un examen approfondi du rôle des images durant l’improvisation
organisationnelle.
La centralisation des données permet la traçabilité des emails. Or la traçabilité a contribué au
genre de communication dépassionné. Nos résultats empiriques nous incitent donc à faire
preuve de prudence à l’égard de cette propriété technologique. D’un certain côté, la
centralisation des données permet une narration et le développement d’un référentiel commun
indispensable à l’improvisation. D’un autre, elle peut impliquer une certaine appréhension ou

  51
frilosité des acteurs à l’idée de discuter de problèmes, de partager leurs expériences ou de
prendre le risque de soumettre de nouvelles idées.
Les emails n’incluent pas une puissance de calcul, mais permettent de récolter des données
afin d’alimenter des bases de données et des algorithmes de calcul. En ce sens, la technologie
de l’email peut être associée à cette propriété technologique. Nos résultats empiriques nous
conduisent toutefois à nuancer nos propos concernant l’effet de la puissance de calcul sur
l’improvisation organisationnelle. La puissance de calcul implique des transmissions
fréquentes de mesures entre les acteurs de la crise. Or, en privilégiant la communication
autour de chiffres, l’utilisation de la technologie peut entraver l’échange d’informations
informelles, non chiffrées, d’opinions et d’émotions. Le phénomène de chasse à la donnée qui
a caractérisé le genre de communication dépassionné durant la canicule de 2003 illustre le
risque d’une attention disproportionnée aux données, laissant de connecter des occasions de
discussion entre les interlocuteurs.
En fonction du genre de communication que les acteurs ont développé autour de la
technologie, la virtualité, la centralisation des données et la communication multilatérale
peuvent représenter des outils de contrôle. En effet, les acteurs de la réponse à la canicule de
2003 ont perçu ces propriétés comme une potentielle menace professionnelle, ce qui les a
conduits à privilégier durant la communication électronique la narration des processus afin de
se protéger d’éventuelles désapprobations. Les calculs stratégiques des acteurs et le genre de
communication se sont cristallisés autour de cette propriété, limitant la mise en confiance des
acteurs, leur spontanéité et l’expression dialogique autour de l’improvisation.
Enfin, la virtualité permet de recopier et transmettre un contenu sans coût et sans délai
(Schultze, Orlikowski, 2001). Elle renforce également le caractère intangible de l’information
dans la mesure où il n’y pas a de lien matériel évident entre l’information et l’objet qui le
contient ou le transmet. Le volet théorique de la thèse a présenté l’apport potentiel de la
virtualité à l’improvisation en situation de crise. Par une transmission sans effort, les
interactions sont plus nombreuses et permettent un partage harmonieux des informations et
des connaissances. La discussion est également fluidifiée, ce qui facilite l’interaction entre les
experts. En facilitant la transmission de la voix et des images, la technologie permet
également une plus grande richesse de l’interaction (Fitrianie et al., 2007; Thomas et al.,
2009).
Toutefois, la virtualité implique l’utilisation d’une même interface pour communiquer,
limitant ainsi l’utilisation de supports de communications supplémentaires comme du papier.
Durant la réponse à la canicule de 2003, les utilisateurs des emails ont rarement utilisé
d’autres médias de communication. Or la matérialité d’un média, comme le fax, permet la
représentation parfois innovante de l’information sur de nouveaux supports. Par exemple
l’utilisation du papier a favorisé l’intégration manuscrite d’avis personnels ou de nouvelles
idées, signe de l’adaptation du genre de communication dépassionné. En somme les médias
les plus « matériels » ont été l’objet d’improvisation. Ce résultat reflète l’importance du rôle
des objets dans l’improvisation, comme le suggèrent les travaux de Claude Lévi-Strauss sur le
bricolage (1962).
  52
La Figure 12 représente les liens conceptuels entre la technologie, le genre de communication
dépassionné et l’improvisation organisationnelle.

Figure 12. Confrontation des résultats théoriques et empiriques

  53
 4.III. Implications managériales de nos résultats
La thèse s’inscrit dans une approche pragmatique et réaliste. Il s’ensuit la nécessité de
formuler des recommandations concrètes à destination des praticiens.
Notre première suggestion est une plus ample reconnaissance de l’improvisation dans les
organisations. La difficulté du management de l’improvisation est qu’on ne peut pas prévoir
comment, avec qui, et quand l’improvisation va se dérouler. Il est donc nécessaire que les
managers reconnaissent l’improvisation in situ. Ils peuvent s’appuyer sur la grille d’analyse
que nous proposons dans cette thèse pour reconnaître dans quelles mesures le flux de travail
dévie par rapport aux routines. Cette démarche nécessite le développement d’une culture
organisationnelle de la communication qui dédramatise les erreurs, ce qui a déjà été
recommandé par de nombreux chercheurs spécialisés dans le domaine de l’improvisation
(Vera, Crossan, 2004). On peut par exemple faire des brainstormings par email afin
d’introduire un nouveau genre de communication.
Enfin, nous suggérons l’ouverture d’un espace de communication propice au développement
de genres de communication compatibles avec l’improvisation dans les organisations. On
peut tout d’abord créer des espaces libres parmi les outils de communication utilisés en
situation de routine, comme un forum sur un Intranet. Par exemple, on peut proposer un
espace de controverse dans lequel les utilisateurs peuvent faire part de leurs opinions. De
même se pose la nécessité d’ouvrir les canaux de communication à un réseau d’interlocuteurs
potentiels, contrairement aux recommandations de Turoff et ses collègues, selon lesquels il
faut structurer et prédéterminer les privilèges de chaque utilisateur des TIC (Turoff et al.,
2004, p. 17). On peut également multiplier les canaux de communication pour diversifier les
genres de communication. Multiplier les canaux de communication implique toutefois un
risque de dispersion et va à l’encontre des recommandations de Jennex (2008) selon lequel les
mises à jour d’information sur la réponse à la crise doivent se faire via un seul média sécurisé.
Nous sommes conscients de la difficulté d’appliquer ces suggestions. D’une manière
générale, les recommandations que nous formulons dans cette thèse impliquent des arbitrages
entre sécurité et liberté. En effet, l’improvisation apparaît dans notre étude comme une
pratique pertinente en situation de crise mais peut semer le trouble dans l’organisation tout en
lésant des parties prenantes. L’ouverture de la culture organisationnelle à l’improvisation
n’est donc bénéfique que si les managers suivent avec attention le déroulement de
l’improvisation. De plus, l’ouverture d’espaces de communication supplémentaires peut peser
sur le respect des contraintes de sécurité de l’information ou mettre en danger la cohérence de
l’action collective.

  54
 Section 5. Contribution et opportunités de recherche

5.I. Contributions de la recherche

Premièrement notre travail propose une explicitation de l’improvisation organisationnelle par
l’analyse des interactions. Nous avons identifié les spécificités des interactions qui composent
l’improvisation organisationnelle lors d’une première analyse du terrain. L’ensemble de nos
propositions théoriques sur l’improvisation est validé par notre étude empirique et d’autres
études (Adrot, Garreau, 2010). Nos résultats permettent également d’enrichir le lien
conceptuel entre les notions de vide organisationnel et l’improvisation organisationnelle.
Enfin, les résultats illustrent également l’importance du rôle des genres de communication
dans le développement de l’improvisation organisationnelle. D’un point de vue théorique nos
résultats enrichissent les études actuelles sur l’improvisation en situation de crise. La
littérature a mis en évidence que la difficulté des acteurs de développer un sens commun de
l’urgence ou du moins une perception commune de la situation de crise (Roux-Dufort,
Vidaillet, 2003a) pouvait compromettre l’émergence de l’improvisation. Nos résultats font
écho à ces travaux et suggèrent que les genres de communication développés dans les
organisations peuvent expliquer la difficulté des acteurs à interagir et développer un sens
commun de l’improvisation.
Deuxièmement, notre recherche vise la compréhension des effets possibles des TIC sur
l’improvisation organisationnelle, qui suscite encore un intérêt restreint dans le domaine des
systèmes d’informations. Nous mettons en évidence que l’impact des TIC sur l’improvisation
organisationnelle diffère d’un outil à un autre. De plus, il semble que l’évolutivité d’un genre
de communication dépend du caractère virtuel ou matériel de la représentation des données
par le média associé au genre de communication. Cependant cette hypothèse nécessite de plus
amples recherches. Enfin nous proposons quelques recommandations à destination des
managers pour gérer l’investissement technologique dédié à la réponse à crise et gérer les
aspects de l’utilisation des TIC liés à l’improvisation de crise.
Enfin, notre travail suggère la nécessité d’enrichir le cadre conceptuel des théories du fit pour
évaluer l’adéquation entre une technologie et les pratiques d’improvisation d’une
organisation. Nous proposons que, pour évaluer l’adéquation d’une technologie, il faut
prendre en compte le genre de communication comme caractéristique organisationnelle.
L’opérationnalisation d’un tel construit requiert de plus amples recherches. Nos résultats
suggèrent également qu’un genre de communication et l’utilisation d’une technologie peuvent
évoluer au fil d’une crise. Nous en déduisons la nécessité d’une approche dynamique du fit
technologique en fonction du développement de l’improvisation.
5.II. Quelques pistes de recherche
Tout d’abord, certains liens entre les TIC et l’improvisation organisationnelle en situation de
crise nécessitent de plus amples recherches. Notre étude théorique ne nous a pas permis de
mettre en évidence comment la technologie peut empêcher l’improvisation organisationnelle
car plus d’éléments positifs que négatifs ressortaient des articles de l’échantillon. De même,

  55
notre étude empirique se concentre sur les limites du genre de communication dépassionné
sur l’improvisation organisationnelle en situation de crise. On pourrait par exemple imaginer
que ce dernier permet aux acteurs de garder leur calme et de ne pas tomber dans l’écueil
d’actions désespérées ou de la panique (Starbuck et al., 1978).
De plus amples recherches permettraient de clarifier le lien entre le vide organisationnel et les
genres de communication, ainsi l’effet de l’improvisation sur les genres de communication.
Une étude approfondissant l’influence de l’improvisation sur les pratiques et les structures
organisationnelle sur le long terme permettrait de faire le lien entre les notions ici explorées et
d’autres concepts tels que l’innovation ou l’apprentissage organisationnel.

  56
 Research questions and document overview
Crisis has increasingly become a subject of extensive research for several years (Boin,
Lagadec, 2000; Dayton, 2004; Pearson, Claire, 1998) and a major concern for managers, due
to increasing organizational and technical complexity (Perrow, 1984) and the emergence of
new unavoidable threats (Perrow, 2006). The contribution of ICT to crisis response has been
discussed in an extensive manner for decades.
One can easily find vivid illustrations of the increasing ubiquity of technology in crisis
response, which makes it necessary to refine our knowledge on ICT support to crisis
response. Since the very beginning of the 21st century, ICTs have become accessible, on
occasion free and easy-to-use means to communicate and interact during crisis response.
Blogs that provide recommendations on how to participate in hurricane social networks2
illustrate that point. As another example, Google Maps was extendedly used during the 2009
swine flu pandemic to mark and share knowledge about victims all over the world. The 2010
Haiti earthquake provided a vivid picture of technology’s potential contribution to crisis
response. As Daniel Stauffacher reports on his blog in 20103, platforms including SMS, Web
and Mail content were leveraged just after the earthquake struck. The extensive use of social
media and Web 2.0 demonstrates how the Internet gathers communities of citizens whose
actions can significantly impact crisis response.
Undeniably planned procedures are useful during crisis response (Pearson et al., 1997) but
should be adjusted as necessary because every crisis is unique and prevention can never be
completely adequate to the situation (Billings et al., 1980; Waugh, Streib, 2006). More than
ever, improvisation is a core dimension of crisis response (Joffre et al., 2006; Roux-Dufort,
Vidaillet, 2003b; Weick, 1993). There is no doubt that crisis response becomes more complex
as it involves a wider range of means and actors whose action potentially deviates from the
framework of established crisis response plans and procedures response (Kreps, Bosworth,
1993; Liu et al., 2008; Palen et al., 2007). In addition, technology significantly influences
actors’ tendency to improvise (Moorman, Miner, 1998a). Given such an evolution, further
knowledge is required on improvisation in crisis response and its connection with technology.
Thanks to a substantive body of work, researchers have documented the cognitive and
behavioral aspects of improvisation (Mendonça, Wallace, 2007; Webb, 2004; Weick, 1993).
They have also proposed integrative definitions and conceptualizations of (Cunha et al., 2003;
Miner et al., 2001; Moorman, Miner, 1998b). However, researchers can experience
difficulties to cope with diversity in research on improvisation. First of all, improvisation has
been studied in radically diverse settings, making it difficult to establish conceptual
connections between findings. Additionally, the mystic aura of improvisation due to its

2
http://inventorspot.com/articles/ten_social_media_methods_fight_hurricane_season_31574
3
Daniel Stauffacher is Chairman of ICT4Peace Foundation. To read his post, please copy and paste
this link : http://www.regards-citoyens.com/article-haiti-and-beyond-getting-it-right-in-crisis-
information-management-by-daniel-stauffacher-49182141.html
  57
artistic echo (Mendonça et al., 2010), contributes to the coexistence of divergent
interpretations of the concept.
Regardless of the divergences between authors’ understandings of improvisation, specific
questions remain unanswered about crisis improvisation. Mendonça, Webb and Butts (2010),
who provide a comprehensive review of the various aspects of improvisation in relation to
crisis response posit the necessity to shed more light on the processes involved in crisis
improvisation, such as crisis responders’ interactions. Given the extemporaneity of
improvisation, how do actors find time to adjust to each other? What information do they
exchange to coordinate effectively? Are there critical factors of success regarding the
interactions that compose organizational improvisation? Literature does provide insights on
these questions (Chelariu et al., 2002; Kamoche, Cunha, 2001; Weick, 1993, 2006; Weick,
Roberts, 1993). However, researchers miss an integrative explanation of how crisis
responders interact and use ICT when improvisation is needed. This dissertation is therefore
designed to answer the following primary research question:
What support does Information and Communication Technology ((ICT)) offer organizational
improvisation during crisis response?
To respond to the primary research question, we address the four following questions:
I. How to cope with diversity in research on improvisation?
II. What possibilities does ICT offer for organizational improvisation in crisis response?
III. Does ICT fit users’ needs with respect to organizational improvisation during crisis
response?
IV. How do users benefit (or not) from ICT when they improvise collectively in crisis
response?
To respond to these research questions, we completed three studies. We address research
question I in the first study. We rely on statistical techniques to confirm that research on
improvisation is diverse. In addition we identify divergent academic perspectives on
improvisation, which can help researchers to understand similarities and dissimilarities
between studies on improvisation. We then respond to research question II from a theoretical
standpoint. We suggest some ICT properties that potentially support organizational
improvisation to occur in crisis response. Finally, we focus on a narrower aspect of
technology and develop an in-depth analysis to respond to questions III and IV from an
empirical perspective. We adopt a practice-based view of technology, as advocated by
Orlikowski (1996), and focus on crisis responders’ use of four technologies - the Internet,
emails, faxes and phones – during the 2003 French heat wave crisis. We then put into
perspective our theoretical and empirical findings to propose a final framework.
This document is divided into four parts. First we make a general presentation of our project
(part 1). We then define the concepts we use and make some theoretical propositions (part
2). Then we present our empirical work (part 3). In the final part of the dissertation is our

  58
conclusion. We present our final findings, discuss the implications of our work and propose
new research perspectives (part 4).
The introductory part (part 1) presents our project globally. We first argue our motivation
for this research (chapter 1.1). We then argue our choice for interaction as our unit of
analysis, detail the research questions that direct our work and present an overview of the
dissertation structure (chapter 1.2). Finally, we present our rationale for following pragmatist
and realist principles in this work (chapter 1.3).
The theoretical part (part 2) presents our conceptual effort that was necessary to provide a
minimum of clarity to our work. First, we present some basic theoretical assumptions we
extracted from literature about crisis and crisis response in large-scale disasters (chapter 2.1).
We then tackle diversity in research on improvisation so as to propose that 8 different
academic perspectives on improvisation (chapter 2.2). The final chapter of the second part
aims to research question II (Chapter 2.3). In this chapter, we first examine literature on
improvisation to make theoretical propositions about the role that interactions play in
organizational improvisation during crisis response. On the basis of these propositions, we
review literature so as to uncover 6 ICT properties that could potentially support
organizational improvisation in crisis response.
In the empirical part (part 3) we propose to go beyond the limits of our theoretical findings.
We first describe the 2003 French heat wave crisis case and its response network (Chapter
3.1). Furthermore, we propose a research design to lead a qualitative and retrospective case
study of the 2003 French heat wave crisis. (Chapter 3.2) presents our research design.
Finally, in (Chapter 3.3) we respond to research questions III and IV by proposing an
integrative framework in relation to ICT use during the 2003 French heat wave crisis
response.
The final part of the dissertation (part 4) aims to put in perspective our theoretical and
empirical findings. We first summarize the findings of each study. We then reflect on these
findings and present avenues for future research.

  59
 Part 1. The dissertation basic features: Rationale,
structure, ontology
This part is an introduction to our work and presents its basic features, the raison d’être and
the structure of the dissertation, as well as the type of knowledge we intend to produce. This
part is structured as follows:
Chapter 1.1 details our rationale for this dissertation work. The center of our interest can be
represented as the union of three concepts: crisis response, organizational improvisation and
communication technologies. Our research addresses ten managerial and theoretical issues
that are related to each intersection between two or three of these concepts. In this chapter, we
detail each concern as a motivation for our research.
Chapter 1.2 provides an overview of the structure of the dissertation. The dissertation is also
articulated in three studies that address four research questions. Each study includes
theoretical or empirical findings that we view valuable enough for publication.
Chapter 1.3 presents our ontological and epistemological position. We adopt a pragmatist
approach to research in general. By doing so, we assume that the value of knowledge resides
in its capacity to change practices. We therefore adopt a realist ontology and epistemology.

  60
Chapter 1.1 Our rationale for this research

Introduction

This chapter aims at explaining our motivation to investigate the primary research question.
We present the theoretical challenges and managerial needs that motivate our work.
The primary research question, entitled “What support do ICTs bring to organizational
improvisation in crisis response?” corresponds to the union of three main concepts: crisis
response, communication technologies and organizational improvisation. We primarily based
the formulation of our motivation for this research project on some exploration of the
literature and informal exchanges with experts.
By doing so, we identify managerial and theoretical concerns that our research is likely to
address. We present them in the following order: from the most general to the most specific.
For instance, we start by examining the need for further research on the concepts of crisis and
improvisation. We then present more specific issues in relation to the interconnections
between the three core themes of our research that we represent in Figure 13: crisis response,
communication technologies and organizational improvisation.

Figure 13. Organizational crisis, ICT, organizational improvisation

Our rationale for addressing managerial and theoretical challenges related to the primary
research questions and its core concepts is twofold. First of all, it comes from our pragmatist
posture (see chapter 1.3). Dewey (1908) argues that the relevance of scientific activities
depends on its practical significance and therefore recommends anticipating how knowledge

  61
production is likely to influence practice. Secondly, this reflection settles the boundaries of
our research interest, which is an essential preliminary to any research (Mason, 2002).
The section is structured as described in the following paragraph and reported in Figure 14.
First of all, we examine why additional research on crisis is needed (section 1). Secondly, we
focus on ICT use during crisis response (section 2) and on the theoretical challenges related to
organizational improvisation (section 3), particularly in the context of crisis (section 4). Then
we will present the factors of a growing attention on the ties between improvisation and
information technology from the MIS research community (section 5). Finally, we will justify
our focus for the dissertation topic: organizational improvisation and ICT use for crisis
response (section 6).

Figure 14. Managerial and theoretical motivations

Section 1. Needs for further research on crisis response

Growing environmental uncertainty and technological complexity make crisis more likely to
occur (Perrow, 1984, 2006; Robert, Lajtha, 2002; Shrivastava, 1988). In the nineties,
Richardson (1994) was already alerting that our society was “increasingly becoming crisis-

  62
ridden”. Similarly, Rosenthal and Kouzmine (1993) reported a governmental declaration that
nations were “at risk”. As a result, more attention to crisis preparation and crisis management
is now required in organizations (Barnett, Pratt, 2000).
It is safe to assert that research on crisis is promoting new managerial practices and that its
practical significance is developing. However research on crisis is facing major challenges
(Quarantelli, 2000). From the theoretical perspective, the ongoing debates attest to the need
for further research on the concept of crisis (Roux-Dufort, 2007). Research communities fall
short of a unified conceptualization of crisis in spite of the considerable amount of work
(Pauchant, Douville, 1993; Shrivastava, 1988). Futhermore new practical vulnerabilities
emerge. Metropolitan areas are increasingly impacted and urban populations are not well
suited for coping with disasters (Quarantelli, 2000). This means that practitioners are still
lacking resources, techniques and recommendations to face crisis management challenges
(Robert, Lajtha, 2002). Researchers need to address these challenges by exploring overlooked
dimensions of crisis response, such as improvisation.

Section 2. Some uncertainties about ICTs support to crisis response

ICT support to crisis response has been discussed already. In spite of extensive evidence that
virtual communication is a necessity in crisis response, ICTs have been criticized for
burdening coordination and information transmission (Comfort, 1993; Quarantelli, 1997)
(Katz, Kahn, 1966; Quarantelli, 1988; Ren et al., 2008). Therefore one can question the
relevance of massive investing in infrastructures devoted to respond to turbulences and
disruptions (Lin, Carley, 1997).
The crisis management literature reports at least six organizational functions, e.g. set of
processes that rely on ICT use during crisis response. ICT allows crisis responders to build
communicational bridges, by locating and tracking crisis responders on the field and
providing updated information. For example, Geographic Information Systems (GIS)
instantaneously provide information from distant locations. Content Management Systems are
also extensively used to keep track of the experts and the other persons who should be
contacted, which supports information and knowledge sharing. Recently, ad hoc usage of the
Internet has enabled citizens to take part in crisis response and to collaborate by exchanging
information, comments, photos and more (Liu et al., 2008; Majchrzak et al., 2007; Palen et
al., 2007). In addition, researchers are beginning to recognize the importance of researching
on-line activities for decision making, including studies on information seeking behavior
during the 9/11 attacks and Hurricane Katrina Blogs (Palen et al., 2007; Palen, Liu, 2007b).
Finally, ICTs help to resolve the problem of data accuracy that usually slows down crisis
response (Chan et al., 2004). They also maintain continuous and relevant contact by
information sharing between groups who are only available virtually (Calloway, Keen, 1996;
Mathew, 2005; Mork, 2002). Then, data centralization enables organization to keep in touch
with experts, which helps to avoid over-simplistic decisions that could be disastrous (Carrel,
2005). Table 1 presents some organizational functions and ICT tools that are commonly used

  63
in crisis response and Annex 1 provides some examples of ICT tools and undergoing projects
related to these functions in crisis response.

Table 1. ICTs in crisis response

Function Information Technologies Most recent authors

Location and Geographic Information Systems Comfort, Sungu, Johnson,

tracking Dunn (2001), Dawes et al.,
(2004).
Geo-Collaborative tools (Geo- Cai, MacEachren, Brewer,
CSCW) McNeese, Sharma, Fuhrmann,
(2005).
Risk identification Threat assessment and Van de Walle, Rutkowski
and management management tools (2006)
Simulation tools Mendonça, Wallace (2004)
Coordinating and Multimodal data fusion tools Fitrianie and colleagues (2007)
feedback Common operating picture tools Landgren (2007)
Internet response grid Palen, Liu (2007a), Palen,
Hiltz, Liu (2007), Jaeger et al.
(2007), Majchrzak, Jarvenpaa,
Hollingshead (2007).
Collaborative virtual workplaces Russo, Raposo, Fernando,
Gattass, Karlsson (2006)
Collins, Powell, Dunford,
Mane, Martinez (2008).
Adaptive ontologies (Malizia et al., 2005) (Segev,
2008)
Decision making Decision support systems Comfort (1993), Mendonça
during response (Edvardsson et al.) (2007), French Turoff, (2007),
Carver, Turoff (2007).
Intelligent Agents/Systems Yuan, Detlor (2005)
Information and Shared databases Larson, Metzger, Cahn (2006);
knowledge sharing Carver, Turoff (2007).
Content Management Systems & Collins et al. (2008).
Knowledge Systems
Emergency alerting & alert (Van de Walle, Turoff, 2007)
notification tools
External Internet Perry, Taylor, Doerfel (2003)
communication
However, existing discourse about crisis response occasionally conveys a deceptively simple
image of ICT use. A more realistic and integrative explanation of ICT support is still needed
due to this simplistic picture. For example, ICTs have been generally depicted as a means for
organizations to learn how to respond intelligently when threats put organizational
functioning in danger. However, this view assumes the existence of specific skills for an
effective use of technologies in crisis (Nunamaker Jr et al., 1989), which do not always exist

  64
in the real world. Practically speaking, technology is problematic in that some users tend to
rely on it to provide solutions rather than on their own competencies and good sense. In such
cases, technology finally alters the organizational resilience that we corresponds to the
organization’s ability to deal with its vulnerabilities (McManus, 2008).
Other criticism targets technology functionalities and characteristics. In fact, some of them
may obstruct crisis response given users’ practices (Dawes et al., 2004; Fischer Iii, 1996). For
example, many-to-many communication, which is a core functionality of technology
(Quarantelli, 1997), may require that users permanently stay connected to each other during
crisis. Such a constraint may slow down local action (Jaeger et al., 2007). Likewise, this
functionality also creates a dependence on technology that may be risky: lack of data or loss
of access to information can bewilder organizations during a crisis (Hutchins, 1991). Another
criticism targets ICT design. In times of emotional distress and time constraints, technological
tools do not represent information in the most appropriate way. For instance, immediate
feedback from a distant location can serve as a support to coordination during disasters but it
also means actors will have to deal with large amounts of information without enough time to
filter it, jeopardizing crisis response due to information overload (Hiltz, Turoff, 1985). ICT
does not facilitate reflection on alternative inputs and processes that may be necessary to
respond to a crisis (Mendonça et al., 2004a; Quarantelli, 1988; Suchman, 2007). ICTs may
also fail at providing an appropriate representation of information exchanges, which may
puzzle users individuals who have to assume ad hoc responsibilities and tasks in disasters
(Mendonça, 2007).
Support from ICTs to crisis response is not fully explained and therefore uncertain. In
addition, because of our lack of knowledge on how and why technology supports or
constrains crisis response, managers lack guidelines on proper practice. Given the importance
of information in modern organizations, managers need a clear outlook on potential
technological fallouts associated with crisis response. For managers, this implies the need for
practical guidelines to avoid technological pitfalls and to improve information management in
crisis response. From a theoretical standpoint, it promotes more reflection on what makes ICT
an obstacle for crisis response.

Section 3. Organizational improvisation: some theoretical challenges

The concept of improvisation is progressively attracting researchers from various fields such
as MIS, marketing, entrapreneurial activities and human resources. There have been various
approaches to this concept. For example, improvisation has been seen as a cognitive process
that involves spontaneity, creativity and intuition (Vera, Crossan, 2004). Improvisation has
also been conceptualized as problem solving in various circumstances, such as product
innovation (Moorman, Miner, 1998b) and strategy formulation in a high velocity environment
(Brown, Eisenhardt, 1997).
Researchers have led great efforts to define improvisation. Cunha, Cunha and Kamoche have
classified more than one hundred papers that use the concept of improvisation (1999c).

  65
Similarly Moorman and Miner shed light on the concept of organizational improvisation by
defining a measurement scale of organizational improvisation (1998a). Mendonça and
Wallace, on their side, have provided theoretical foundations for computerized simulation and
support to the cognitive phenomena that entails improvisation (2007).
Although these authors have provided significant insights on organizational improvisation,
they fall short of details on the dynamics underlying its development. By relying on an
incomplete conceptualization of improvisation researchers risk providing over-simplified and
unreliable recommendations to managers, impeding organizational efforts to improvise
successfully (Gauthereau, Hollnagel, 2005; Vera, Crossan, 2005). More information is needed
on individuals’ participating in improvisation. In particular, interaction is a crucial dimension
of organizational improvisation that requires further investigation. Both the content and the
nature of interactions between improvisers remain unclear. The literature on crisis response
has partially addressed this issue. Weick (1993) and Ciborra (1996b) present the Mann Gulch
disaster as a vivid example of the irreversible consequences of miscommunication during
improvisation. As everyone’s astonishment when Dodge lit a fire illustrates, improvisers can
disconcert their colleagues by their apparent behavior, all the more in critical situations where
emotional and time pressures are at their climax. Researchers have provided additional
insights on individuals’ ability to share their intuition and moods (Ciborra, 1999) and to
develop a collective mind (Weick, Roberts, 1993). However, theories that mention the
importance of interactions between improvisers fail at opening the “black box” of interactions
and at providing an integrative view on interactions and their role in improvisation. The need
to determine what interactions do or do not occur during improvisation and whether
interactions are specific or similar to any other process in organizations still exists.
Finally, improvisation remains conceptually challenging in that it provides divergent
meanings. Even though some authors view improvisation as an alternative to planning, some
others conceive it as a complementary approach. For this reason, further study is required on
improvisation as not only an interactive process, but also as a polysemous concept.

  66
 Section 4. The rationale for studying improvisation in crisis response

From a practical perspective, it is necessary for managers to prepare for crisis improvisation.
Contrasting with this need, managers are reluctant to mindfully monitor improvisation or even
admit improvisation might occur within their organization. Most of the time, such a lack of
awareness makes the perception of improvisation retrospective (Weick, 1993). Consequently,
it prevents managers from effectively reacting to the ongoing improvisation in crisis response.
From a theoretical perspective, studying improvisation in crisis settings is an opportunity to
develop knowledge on improvisation. As Roux-Dufort explains (2007), crisis exposes
organizations to extreme situations in which uncertainty, emotional pressure and time
constraints are higher, which provides the opportunity to observe some organizational
phenomena more easily: “In organizational studies, crisis usually serves as an amplification
chamber for research” (Starbuck, Farjoun, 2002, cited by Roux-Dufort, 2007). That being
said, one cannot directly extrapolate the findings from an observation of a specific crisis
response to any routine context.
While, researchers have commonly admitted improvisation refers to a unique phenomenon,
they also depict crisis improvisation and routine improvisation as similar phenomena. In crisis
response, improvisation seems to be triggered by contingency and actors are compelled to
improvise. Conversely, actors seem to have the choice to improvise in routine. Improvisation
emerges as an endogenous deviation from established patterns of actions. However the
specificities of crisis improvisation require further investigation.
Finally, it is necessary to discern the benefits of improvisational behavior from its fallouts.
The Mann Gulch disaster illustrates how a single improvisation can have both positive and
negative outcomes. Dodge’s survival proves that actors can experience improvisation in an
overall positive manner. However Dodge’s colleagues were disturbed by his improvisation
and were killed by the fire. This dissertation addresses the first of the conceptual ties
mentioned above: the factors of improvisation in crisis response.

Section 5. The IS researchers’ perspective on improvisation

Organizations’ likelihood to improvise increases with information flows (Chelariu et al.,

2002). However research on ICT and improvisation remains scarce.
In the last decades, there have been calls for more attention in the MIS field on unexplored
concepts such as improvisation. Nowadays, researchers have been increasingly turning to
concepts of emergence and improvisation to explain new ways of organizing and using
technology (Orlikowski, 1996; Orlikowski, Hofman, 1997). However the fundamental role of
technology in daily reality, such as in Internet use, has been overlooked. Researchers have
focused on system methodologies and designs such as CASE tools, BPR and CMM because
they are easier to conceptualize (Ciborra, Hanseth, 1998). On the contrary, daily life is full of
bricolage, improvisation, or serendipity (Latour, 2004). Even though these themes have been
neglected, they are core components of individual and collective action and therefore highly
  67
significant for research in MIS fields. There is the need for IS research to focus its attention
on notions such as improvisation even if these notions are difficult to conceptualize and
observe.
At the same time, organizations are currently experiencing changes that make improvisation
an even more relevant topic. Improvisation is increasingly necessary for managers because of
a higher level of environmental complexity, uncertainty (Vera, Crossan, 2005) and velocity
(Brown, Eisenhardt, 1997). The widespread access to technology has created additional
opportunities for improvisation (Moorman, Miner, 1998a). First of all, organizations have to
improvise when technological designs are inadequate to their needs and processes. Users are
likely to develop new practices to adapt to information systems that do not meet their needs
(Heeks, 2002b). Secondly, by providing direct access to informational resources, technology
provides more autonomy for decision makers and encourages them to develop ideas more
extemporaneously (McKnight, Bontis, 2002; Moorman, Miner, 1998a). However, we wonder
whether technology generates improvisation or not.
Organizations are more likely to improvise and information software development should
take this trend into account (Dyba, 2000). Because ICT projects are an important investment,
the lack of support from technology to improvisation is problematic. Managers need to
evaluate how much a certain infrastructure is likely to support improvisation. This requires a
better understanding of the possible tensions between the need to improvise and ICT
characteristics, and also of how people use ICTs when they improvise. Understanding why
and how people improvise when using technology will also help to reduce the risk of systems
failures in the future (Heeks, 2006). This understanding is all the more applicable when the
players and the stakes are human, like in crisis. In this dissertation we address the questions
related to the dynamics of ICT use and improvisation.

Section 6. The Dissertation topic: ICTs and organizational

improvisation during crisis

We have identified four key issues related to the connection between technology and
improvisation in crisis response.
First of all, there is a developing reflection on ICT requirements for collective improvisation
during crisis. How should we design our technology to better support improvisation in crisis
response? It is true that the use of inadequate tools may have disastrous effects on crisis
response. There are few authors addressing this issue. Mendonça (2007) emphasized the lack
of technological fit to the need for improvisation in emergency response, which currently
leads to some reflection on ICT design. Managers need adequate interfaces to remain resilient
in crisis response.
This leads us to the second question with respect to ICTs and crisis improvisation. How do
improvisers use technologies and do they use them in a beneficial way for improvisation?
This question is less normative than the first but commonly reveals the need for analyzing

  68
how people concretely use ICT during improvisation. We address this point within this
dissertation.
The third question focuses on improvisation that is required due to the lack of fit between
ICTs and improvisation in crisis response. Orlikowski (1996) developed her first thoughts on
this subject in 1996 by acknowledging that users improvise when they use technology.
Dawes, Cresswell, and Cahan (2004) explain how crisis responders make do without the
appropriate information format. ICT use may imply improvisation and bricolage (Ciborra,
1996b; Pawlowski et al., 2004), which is particularly true in crisis situations. Managers need
recommendations to guide users in their improvised use of technology.
Finally, the fourth question addresses the influence of ICT on improvisation. How does using
technology influence crisis response? In particular, does it affect the odds for improvisation?
Does it affect the probability of success for improvisation? Is improvisation more frequent in
organizations that have access to ICT for crisis response? So far, there is no evidence of how
ICT organizational improvisation in critical situations. Responding to that question requires a
preliminary analysis of the nature of the use of ICT during organizational improvisation for
crisis response. We directly address this issue in this dissertation.

  69
Conclusion
Table 2 gathers the key issues that we have identified so far. Each line summarizes key issues,
pending managerial needs and theoretical challenges that we detailed in each section. We
intend to respond to some of the listed key issues. Others are only partially addressed or not
addressed at all.

Table 2. Key issues that we address (or not) in the dissertation

Key Issue Managerial needs Theoretical Challenges Is it addressed

by our work?

Crises become To multiply potential 1. The need for a unified Managerial

more and more techniques and definition of crisis. needs: Yes, by
frequent resources to face crises 2. The need to develop our providing
(Section 1) (Perrow, 2006) knowledge on some guidelines and
unexplored aspects of crisis definitions of
response such as crisis
improvisation. improvisation

Theoretical
challenge 1: No

Theoretical
challenge 2: Yes.

ICT can 1. To evaluate the The need to report what Managerial Need
burden crisis relevance of ICT makes ICT become an 1: Partially, by
response or at investment to prepare obstacle to crisis response identifying ICT
least reduce for crisis response characteristics
organizational 2. To develop practices that are likely to
resilience to properly manage influence
(Section 2) ICT and avoid organizational
technological pitfalls improvisation
during crisis response during crisis
response.

Managerial need
2: Partially, by
identifying
patterns of ICT
use during crisis
response (see part
3).

  70
 Theoretical
challenge: Yes,
by analyzing ICT
use dynamics and
effects on crisis
improvisation.

Managers Guidelines about how 1. There is the challenge Manager needs:

remain to figure and manage address the limitations of the Yes.
skeptical about improvisation. theories on improvisation
improvisation Theoretical
2. The need to develop an
(Section 3). challenge 1: Yes,
integrative view of
by studying
interactions that occur during
improvisation in
organizational improvisation
critical settings.

Theoretical
challenge 2: Yes,
by studying
improvisation in
critical settings.

There is no A representation of 1. There is little explanation Manager needs:

evidence of improvisation and its in of the occurrence of No.
improvisation and outs. Managers are improvisation in critical
being reluctant to mindfully settings Theoretical
beneficial or monitor or admit challenge 1:
2. No evidence that crisis
not to crisis improvisation in crisis Partially. We
improvisation is significantly
response response, even if analyze how ICT
similar to routine
(Section 4). improvisation is not a use influences
improvisation
matter of choice. organizational
3. Lack of knowledge on the improvisation
benefits of improvisation. occurrence in
crisis response.

Theoretical
challenge 2: No.

Theoretical
challenge 3: No.

But ICT 1. Guidelines to 1. Understanding how people Managers need 1:

support to manage tensions use ICT when they improvise Partially.
improvisation between improvisation 2. How does ICT use affect
is not requirements and ICT improvisation?
established. infrastructure.
Managers need 2:
(Section 5) 2. The need to evaluate Partially, by
the ability of a defining
technology to support technological
improvisation before characteristics

  71
 investing. that support
organizational
improvisation in
crisis response.

Theoretical need
1: Yes. It is a
core question in
our empirical
investigation

Theoretical need
2: Yes.

There is scarce 1. The need for an 1. What would be a good Managerial need
knowledge on adequate design for design for improvisation in 1: No.
IT and improvisation and crisis response?
improvisation resilience in crisis Managerial need
2. The need to better
in crisis response 2: Yes.
understand the nature of ICT
response. 2. The need for use during organizational Managerial need
(Section 6). recommendations to improvisation 3: Yes.
guide users in their 3. Understanding how ICT
improvised use of ICT use affects improvisation Theoretical
3. The need for challenge 1: No
4. The need to assess ICT fit
guidelines to use ICT to crisis improvisation
in a beneficial way Theoretical
when improvisation is challenge 2: Yes.
required
Theoretical
challenge 3: Yes.

  72
Annex
In this annex we present an overview of some ICT tools that are commonly used or developed
in crisis management. Numerous ICT projects are being developed. For this reason, we
describe 3 projects that represent a infinitely small proportion of the ongoing work on ICT
and crisis response.
Ushahidi (http://www.ushahidi.com/) is a web interface that enables citizens to view and
exchange information about crisis situations in diverse countries through sms, email or
webforms. It was initially developed to gather information after the post-election violence in
Kenya but it is now used in other countries such as Congo, South Africa, and Mexico.
Ushahidi relies on graphical representation of information with colours to distinguish sources
of danger.

Figure 15. Ushaidi Screenshot

  73
Ushahidi enables users to send alerts and receive alerts, which guarrantees communication
between persons in need and those who detain real-time information. The interface to send
alerts is presented in Figure 15.
INSTEDD Geo Chat enables mobile phone users to exchange information when the situation
is critical. Therefore, crisis responders coordinate with their colleagues by identifying the
other users’ location and activities. InSTEDD functionalities are represented in Figure 16.

Figure 16. InSTEDD screenshot

OCHA enables crisis managers and watchers to exchange all sorts of documents. It targets
humanitarian organizations and aims to support communication and collaboration between
national and international actors to "alleviate human suffering in disasters and emergencies"
as well "facilitate sustainable solutions"(http://ochaonline.un.org). OCHA is represented in
Figure 17.

  74
 Figure 17. OCHA Screenshot

  75
Chapter 1.2 Four questions and three papers: Our rationale for the
structure of the dissertation

Introduction

As a reminder, we started our dissertation project with the following research question in
mind: “What support does ICT offer to organizational improvisation in crisis response?”
Because this research question is broad, we clarify the three topics that we address in the
dissertation: i) crisis response, ii) organizational improvisation and iii) Information and
Communication Technologies (ICTs).
First of all, divergent conceptualizations of organizational crisis and crisis response coexist.
Therefore we detail our understanding of these concepts and we specify the type of crisis
response we investigate in chapter 2.1.
Secondly, the management literature on improvisation appears diverse at first glance. Authors
report various examples improvisation. In addition, they analyze improvisation from distinct
perspectives. However, scholars who investigate the concept of improvisation can experience
difficulties in making sense of such diversity and in contributing to knowledge building on
this concept. As a result, our investigation on ICT support to organizational improvisation in
crisis response requires that we preliminary address research question I entitled “How to cope
with diversity in research on improvisation?” in chapter 2.2.
Finally, Orlikowski and Iacono claim the need to reflect on what ICT consists of when
starting a research in the MIS field (2001). First of all, the lack of definition of technology
threats the relevance of MIS research (Orlikowski, Iacono, 2001) by making researchers miss
crucial details for a thorough analysis (Latour, 2004). Secondly, technology is a single term
that covers a plethora of tools and usages, from the use of databases for transversal
coordination to the use of Web 2.0 and social media. A quick glance at the latest disasters
unveils that both improvisation and ICT use can refer to distinct situations. For instance,
Iranian citizens spontaneously used twitter to deliver fresh street level information during
mass protest in June 20094. In the case of Twitter use during the 2009 Iranian mass protest,
improvisation related to an unexpected and innovative use of an ICT tool. The international
response in the aftermath of the 2010 Haiti earthquake5 provides another example of ICT use
for crisis improvisation. As described by journalists, crisis responders needed ICT to
communicate their needs to manage improvised ad hoc solutions. This case significantly
differs from Twitter use in Iran: Technology was not the object of improvisation but
supported coordination between improvisers. As a journalist described:
“One member of the International Federation of Red Cross (IFRC) assessment teams stopped
at several to eyeball the situation and talk with the locals. Clipboard in hand, he jotted down

4
http://www.time.com/time/world/article/0,8599,1905125,00.html
5
http://www.wired.com/magazine/2010/04/ff_haiti/all/1
  76
key indicators like access to water, numbers of children, availability of improvised shelter
materials, and whether any other aid outfits had already been there. He snapped the
occasional digital photo to augment the reports and logged the location of each camp with a
GPS unit — critical in places like Haiti where there never were many street addresses to
begin with. Back at base camp, his findings would be added to those of other assessment
teams, along with information from media reports and other sources, in an ever-growing
database.”
Therefore, the objective of this chapter is twofold. First we present the four research questions
that were developed from our primary research question e.g. the generic question of
technology support to organizational improvisation. Second we explain how each study
responds to the research questions. By doing so, we detail how the dissertation project is
articulated.
This chapter is structured as follows: In Section 1 defines the dimensions of ICT and ICT
support that we retain in this dissertation project. In Section 2 we argue the relevance of using
interaction as a unit of analysis to lead a thorough examination of crisis improvisation and its
linkages with technology. In Section 3 we present an overview of the dissertation questions
and how they relate to each other. Finally Section 4 presents each of the three studies and how
they address the four research questions.

Section 1. Our rationale for the dissertation structure

1.I. Three perspectives on ICT

We identify three perspectives on ICT in the literature. We rely on these perspectives to
respond to the primary research question.
At first glance, ICT can be examined as a collection of socially recognized properties.
Orlikowski and Iacono define the technology artifact in the following manner (2001): “those
bundles of material and cultural properties packaged in some socially recognizable from such
as hardware and/or software”. In this vein, some studies focus on the common features of
ICTs and their impacts on organizations. In this kind of work, authors develop an overall
analysis of ICT. Not only do these authors adopt a generic view of ICTs but they also analyze
the influence of technology at the organizational or societal level. For instance, Henderson
and Venkatraman argue that organizations need to adapt their ICT infrastructures to their
strategies to increase performance (1993). Similarly significant amount of studies explore the
mechanisms by which technology supports the development of new capabilities (Pavlou et
al., 2004). A generic view of ICT impact on organizations exists in crisis management
literature as well. For instance, Quarantelli questions the problematic aspects of technology at
a general level (1997, 2007). He explains how easy replicating and exchanging information
are, therefore provoking information overloads that can burden crisis response.
A generic view of technology is valuable to discuss the relevance of ICT investments and
implementation. However, including the users’ contingency into analysis enables researchers
to expand their scope of inquiry to some practical dimensions of ICT (Orlikowski, 1992). ICT
  77
Simon (1996) defines the artifact as a “meeting point” between an inner environment, “the
substance and the organization of the artifact itself” and an outer environment. Therefore an
alternate approach consists of examining ICT as a set of interaction between a system of
computerized services and the user. This approach aims to assessing whether technology
responds to its users’ needs. A significant amount of theories follows that approach, such as
technological fit theories (Goodhue, Thompson, 1995; Mathieson, Keil, 1998; Rice, 1992)
and the technological affordances (Gaver, 1991). These studies explore ICT at a more specific
level of analysis than the first approach. They usually take into account a specific service ICT
offers, such as risk assessing, decision-making or even communication. In addition, they
focus on a specific set of organizational processes, usually the ones that rely on the ICT
services. For instance, Massey and Montoya-Weiss focus on knowledge conversion in
organizations and how several technological tools participate in knowledge conversion
(2006). A considerably high amount of work in MIS research focuses on specific services that
can be rendered by technology. For instance, Markus and her colleagues designed the Top
Modeler tool to support emergent knowledge processes (Markus et al., 2002). Finally, these
studies take into account the users’ perception of technology. For instance the TAM theory
focuses on the technology easiness of use and its users’ perception of the technology
usefulness (Davis, 1989).
Finally, the practice-based view provides additional insights on the relationship between ICT
and its users. Technology is ever-present in daily life, therefore deeply influencing social
systems (Orlikowski, 2007). For example, widespread daily access to Internet played a crucial
role in the spontaneous development of information and support websites used during the
response to Katrina (Palen et al., 2007). In addition, ICT services are shaped by organizational
processes (DeSanctis, Poole, 1994; Masino, Zamarian, 2003). For instance, Katrina crisis
responders’ customized their use of Google Maps on the basis of their informational needs for
evacuation. By doing so, the crisis responders could benefit from the program, using
information that was useful to rescue victims such as the damages and the number of victims.
In other words, the uses and consequences of ICT emerge unpredictably from complex social
interaction (Markus, Robey, 1988; Orlikowski, Robey, 1991). For this reason, studies that
follow the practice-based view of technology focus on users and the organizational context of
ICT use. To do so, these studies analyze ICT by observing groups of users’ practices in
relation to a specific ICT tools. For instance, the improvisational model of technology
practices emerged from the observation of the daily use of Lotus Notes technology in a
customer support department (Orlikowski, 1996). Table 3 presents the three perspectives on
ICT presented here in and their corresponding levels of analysis.

  78
 Table 3. Focus on technology and users

Perspective Level of analysis Description Example(s) of theories

on ICT
ICTs from a Organizations Researchers focus on Alignment theory
general stance IS infrastructure ICT generic (Henderson, Venkatraman,
ICTs properties and 1993)
evaluate their effects
on users and
organizations
Interactive Set of ICT tools with Researchers take TAM theory (Davis, 1989)
view of ICTs common into consideration IT appropriateness theory
and users characteristics both ICT tools (Zigurs, Buckland, 1998)
Specific set of tasks or characteristics and IT affordance (Gaver,
processes user’s tasks. 1991)
The media richness theory
(Daft et al., 1987)
Practice- Users of a specific ICT Researchers analyze The improvisational model
based view of tool. ICT users’ practices of technology practices
ICT (Orlikowski, Hofman,
1997)
The Adaptive structuration
theory (DeSanctis, Poole,
1994)
1.II. ICT Support in crisis response
Directly applying the three perspectives on technology, we propose three definitions of the
notion of ICT support. Even though the term “support” remains vague, referring to something
that provides assistance or that enables specific processes, existing research refers to specific
aspects of support. First as the possibilities offered by technology for crisis response, as the fit
between technology’s features and crisis responders’ needs, and finally as the actors’
leveraging of technology for improvising. We detail each of them below and in Table 4.
First of all, some authors focus on the technology’s primary function of enabling people to
communicate during crisis response. These authors analyze ICT support by exploring the
overall possibilities offered by technologies to crisis responder. By doing so, they explore the
influence of generic features of ICT on crisis response. For instance, some researchers report
how social media enable crucial communication between distant locations, either between
victims or between responding organization and experts (For-mukwai, 2010).
Secondly, other authors explore the fit between ICT characteristics and crisis responder’s
specific needs regarding their interactions. This approach consists of comparing the design of
a specific service provided ICT with its effective use in crisis response. In other terms, this
approach consists of addressing the following question: Does ICT fit the users’ needs and
tasks?

  79
Finally, a third approach to ICT support focuses on how users practically take advantage of
technology to complete their tasks. In relation to ICT support, this approach addresses the
following question: How do users leverage ICT as a resource to complete organizational
processes?

Table 4. Our understanding of ICT support

Perspective on ICT Focus Definition of ICT support

Generic perspective on ICT
Interactive view of ICT and
users
A practice based approach
Organizations
ICT at large
A service provided by ICT
The needs of a group of users
A group of users
One or several ICT tools
Possibilities offered by ICT
to organizations
There is a good fit between
the design of the services
provided by ICT and the
users’ needs
Users practically take
advantage of ICT to complete
their tasks effectively

Section 2. Interactions as unit of analysis: Our rationale

Interaction is a core component of organizational improvisation but has been overlooked. For
this reason, we argue that interaction is a relevant unit of analysis for our study. Our rationale
is double.
2.I. What is an interaction?
The notion of interaction has been widely used for research instrumentation but has been
scarcely defined. To introduce our rationale, we propose to set the definition of what is an
interaction.
Actually, one needs to explore the cybernetics and the psychology literature to know more
about what is an interaction. In his work on schismogenesis in the “Iatmul” tribe (in New-
Guinea), Bateson (1935) describes how relationships between individuals evolve over time.
Depending on how individual A reacts to individual B and how B reacts to A’s reaction to B,
A and B may enter into conflict. Such a sequence of mutual adjustments to each other’s
communications and actions constitutes an interaction (Watzlawick et al., 1967). Existing
definitions of interaction are consistent with Bateson’s work. For Ruesch and Bateson (1951),
an interaction can be depicted as an intricate sequence of intertwined individual and collective
reactions. Newcomb, Turner and Converse (1965), as well, define interaction as a sequence of
behavioral patterns from individuals who react to each other. In that sense, an interaction can
be observed through a succession of interdependent actions from two or more individuals
(Hare et al., 1965; Newcomb et al., 1970). Such definitions mirror the etymologic origin of
the word. Interaction literally means between (inter) actions (Tyndall, 1867 – etymologic
dictionary). However, as we suggested here in, an interaction includes a great deal of

  80
communication. Rather, interaction can be viewed as a process that includes actions,
reactions, and communication (Bales, 1950).
Communication is a core component of interaction but radically differs in that it consists of an
exchange of information X between two sources A and B (Newcomb, 1953; Ruesch, Bateson,
1951) that may be influenced by the interlocutors’ actions. As Figure 18 illustrates, the
concept of communication corresponds to verbal and non-verbal signs between A and B: A
talks to B but also expresses affects through gesture and face expression. In return, B
transmits some response to A. As Bales (1950) explains, communication is both an ingredient
for and an achievement from interactions.
Finally, interaction is systemic and refers to a complex set of actors and events in time and
space (Watzlawick et al., 1967). Not only do A and B react to each other, but they are also
embedded in a complex set of mutual adjustments (Goffman, 1967, 1970).
In other words, an interaction is situated, which means its observation should not be restricted
to a sum of bilateral exchanges but rather understood in its complexity (Goffman, 1967, 1974;
Watzlawick et al., 1967).
The literature has suggested a general definition of an interaction that refers to a complex and
dynamic set of exchanges between actors and events. Such a broad definition seems correct
enough and highlights how an interaction inherently shapes the course of action. We therefore
define an interindividual interaction as the sequence of communications between several
individuals that enable or make them change their actions accordingly”. Definitions of
interactions remain abstract. For this reason, this research hinges on Bales’s work to identify
interactions (1950). To study, observe and analyze an interaction, we view it necessary to
refer to a specific list of possible interactions, as suggested by Bales. Describing an
interaction as a process, Bales (1950) places the emphasis on the meaning of each
communication that composes interaction. By doing so, he presents 12 categories of
interactions that are divided in four main groups and complementary by pairs. Figure 18
represents these interactions.

  81
 Figure 18. Interaction categories, Bales (1950)

Task area interactions significantly differ from socio-emotional area interactions. While the
former gathers questions and answers related to a specific problem or task that the interaction
is related to, the latter refers to negative and positive emotional and social reactions. Bales’s
categories well match to the context of our research.
Many authors have emphasized the importance of interactions in collective life. As suggested
by the notion of symbolic interactionism (Mead, 1934), individuals adapt their actions on the
basis of their social interactions Furthermore, meanings are handled and modified through an
interactive process (Blumer, 1986). Such a view is consistent with the pragmatist posture
(Strauss, Corbin, 2008) that we follow throughout the dissertation.
Arguments supporting the importance of interactions can be found in French sociology. Not
only do interactions influence individual development but they are a core component of social
life (Piaget, 1977). Finally, research on organization underscores the importance of
interaction. For Weick (2001b), interactions support good coordination. In this sense,
observing interactions is a way to provide rich insight on how actors dynamically understand
and adjust to each other. In the coming section, we will argue the importance of interactions
in crisis response and crisis improvisation.

  82
 2.II.The importance of interactions in crisis response and crisis improvisation
Generally speaking, coordination is not only the result of task programming (Van de Ven
(Faraj, Sproull, 2000; Mintzberg, 1996; Van de Ven et al., 1976) but also of interactions
between co-workers. Interactions, notably information feedback loops (March, Simon, 1958),
enable mutual adjustments (Mintzberg, 1996) especially in critical situations (Dynes,
Quarantelli, 1976). As Espinosa Lerch and Kraut (2002) explain, interactions indirectly
influence coordination by four processes: the emergence of common mental frames (Espinosa
et al., 2002) team situation awareness (Weick, Sutcliffe, 2001), transactive memory (Brandon,
Hollingshead, 2004; Wegner, 1987) and collective mind (Weick, Roberts, 1993).
Depending on how people interact with each other, resources are allocated and distributed.
Thus the perspective of having reliable interactions with collaborators during crisis response
is necessary to maintain trust, good collaboration within the whole organization (Comfort,
2007; Mishra, 1996; Webb, 1996). Effective interactions also prevent misunderstandings that
are frequent when actors involved in crisis response are diverse. From public organizations
and officials, to non-profit organizations and citizens, the response to Katrina included more
than 265 different entities (Comfort, 2007). Effective interaction between actors is also
important because collective mindfulness, resilience and extra-activities are needed during
response (Weick, Sutcliffe, 2001). An interaction is likely to influence people coordinating
with and adjusting to each other, and finally their participating in crisis response (Comfort,
2007).
While effective communication can support crisis response, ineffective communication can
lead to misunderstandings and inaction. For example, in the Mann Gulch Disaster actors
experienced problems interacting, which contributed to their isolation and their decisions
made in despair (Ciborra, 1996b; Weick, 1993). Additional examples include the Bhopal
disaster, during which the incomplete transfer of information between the experts and the
police service led to the death of thousands of people (Shrivastava, 1987). Interactions
become all the more delicate as the number of stakeholders in crisis response increases.
Following the 9/11 attacks, a massive group of citizen volunteers arrived on the location of
the incident, slowing down the emergency units’ work. In spite of their good intentions, the
volunteers required care and preparation and interacted more emotionally than professionally
(Dawes et al., 2004). More interactions have the potential to burden crisis response, which
explains why more managerial recommendations on how to manage interactions during crisis
response are needed.
In the context of crisis improvisation, interactions are more important than ever (Ciborra,
1996b; Montuori, 2003; Weick, 1993). One could argue that improvising is simpler than
struggling to follow procedures. However improvising adds complexity to interaction in crisis
response, especially in relation to interactions. Through interaction, actors construct a
common narrative of past activities (Brown, Duguid, 1991). If deviation from established
patterns of action exists, then the interaction is more likely to either be ambiguous or happen
either too late or too early for synchronization. For instance, nurses were spontaneously

  83
contacted at unscheduled hours during the 2003 French Heat Wave. They had to make the
effort of coordinating accordingly.
2.III. Interaction: a convenient unit of analysis to observe ICT
Finally, human interactions nowadays mirror ICT use (Orlikowski, 2007) and we view them
as a relevant unit of analysis to gain more knowledge on ICT. As some authors point out,
studying technology through human actions and interactions is necessary not only because
technology is increasingly embedded in social life but also because technology essence is
situated – e.g one needs to take into account its context of use to make sense of a technology.
Not only do interactions comprise organizational improvisation, but they also constitute a
palpable reality of ICT use and of the technology artifact. Broadly speaking, people
increasingly rely on communication technologies to interact and for decades interactions have
ceased to be offline exclusively. Progressively, widespread access to ICT nowadays (Garcia et
al., 2009) such as Internet, telephone, instant messaging and email, deeply modifies the
overall context of interaction (Winkin, 2001). As a result, the observation of interactions
through technology enables us to uncover evidence of technology use during organizational
improvisation.

Section 3. Research questions specifications and studies overview

In the previous sections we demarcate possible avenues to address the question of ICT
support in the dissertation. In this section, we refine those propositions to define the research
questions that we address in this work.
For a start, we focus on improvisation polysemy. Indeed there is a loose collection of
representations of this term. Each of them corresponds to a specific epistemological stance
and avenues of research on improvisation. In our first study, we propose to address the
research question I, entitled How to cope with diversity in research on improvisation?
We then focus on ICT. In our second study, we propose to examine the influence of ICT
properties on the occurrence and development of interactions that are intrinsic to crisis
improvisation. By doing so, we address the research question II, entitled “What possibilities
does ICT offer for organizational improvisation in crisis response?”
Given our focus on technology as a distinct entity, differentiating it from user’s practices, we
propose a theoretical investigation.
Viewing ICT fit and ICT use as intricately related concepts, we propose to tackle question III
and IV altogether. Thus we address the two following research question in our third study:
Research question 3: Does ICT fit users’ needs with respect to organizational improvisation
during crisis response?
Research question 4: How do users benefit (or not) from ICT when they improvise
collectively in crisis response?

  84
Findings from various papers are then confronted and discussed to uncover the theoretical and
managerial implications.
3.I. Chapter 2.2. How to cope with polysemous concepts in organization studies?
The case of improvisation
For two decades, improvisation has been attracting the attention of researchers in the
management field. Researchers have defined and studied improvisation in various manners,
strengthening its polysemy. However, scholars lack a classification scheme to evaluate
differences and commonalities between studies on improvisation, which lessens opportunities
for discussion to improve our understanding of improvisation. To handle this apparent
academic diversity that surrounds the concept of improvisation, one needs to assess
dissimilarities and common points among studies. In addition, researchers need a
comprehensive outlook of the dimensions that account for the diversity within research on
improvisation. We responded to this need by using CATPCA and PCA techniques. We also
used hierarchical cluster analysis techniques to propose a representation of the divergences
between existing approaches of improvisation in management literature. Our findings confirm
that research in this field has been eclectic to the point that no predominant view of
improvisation exists. Additionally, we propose that diversity stems from the academic
divergence of authors in relation to four constituents of the research process: i) evidencing, ii)
focalizing, iii) contextualizing, and iv) creating knowledge on improvisation. We finally
discuss the theoretical implications of our findings and future avenues for research.
3.II.Chapter 2.3 Information technology, crisis and organizational improvisation:
Some theoretical propositions and a literature review
The objectives of this paper are twofold. First of all, we propose a new and general definition
of organizational improvisation. We conceptualize organizational improvisation through the
dynamic capabilities approach. The first part of this article consists of a proposal for theory
on organizational improvisation. We rely on previous work to suggest three components of
the interactions that compose organizational improvisation: boundary spanning, expertise
leadership and minimal structures of knowledge.
Then, based upon our suggestions, we conduct a literature review on crisis management. In
our literature review, we retain articles that deal with ICT use and in which we partially or
fully recognize our definition for organizational improvisation. Therefore, our literature
review gathers articles from very diverse fields of research.
Finally, we identify five technological characteristics that are likely to support some of the
components of organizational improvisation: Many-to-many communication, graphical
representation, data centralization, calculation and virtuality.
3.III. Chapter 3.3 ICT, organizational improvisation and emptiness during the
2003 French heat wave crisis response
This paper seeks to understand the role played by Information and Communication
Technologies (ICT) in Organizational Improvisation (OI) during crisis response. The crisis

  85
management literature fails at providing a unified assessment of ICT support to
improvisation. The theories of fit are commonly used to assess whether technology supports
some specific organizational processes. However, they do not fully capture the complexity of
improvisation and crisis response. Due to the lack of theoretical background in relation to ICT
support to crisis improvisation, we conduct a retrospective qualitative analysis of the 2003
French heat wave crisis response in which both administrative and field actors were involved.
Our analysis is based on the grounded theory method and focuses on the interactions that took
place between the crisis responders. We make an integrative framework emerge from data.
This framework suggests that organizational improvisation enabled crisis responders to face
organizational emptiness. However organizational improvisation growth depended on the
crisis responders’ communicative genres. While a fervent communicative genre supported the
development of organizational improvisation among the field actors, a dispassionate
communicative genre, developed by the administrative actors around emails and faxes,
hindered the administrative actors’ participating in organizational improvisation. Finally we
discuss our findings and the contribution of this work.

Conclusion

Figure 19 represents our rationale for building the structure of the dissertation that we depict
in this chapter. The arrows represent that we reflected on conceptual connections between the
propositions and the findings that emerge from our reflection.

  86
 Figure 19. Our rationale for the dissertation structure

  87
Chapter 1.3 Ontological and epistemological Assumptions

Introduction

This chapter presents our ontological and epistemological assumptions. We adopt a

pragmatist, realist approach to explore the question of ICT support to organizational
improvisation in crisis response. This approach has several implications for our research
design that we present in the following sections. Section 1 presents our rationale for being
pragmatists. Section 2 details our ontological assumptions and section 3 considers the
epistemological implications of our posture.

Section 1. A pragmatist posture….

Pragmatist thought was first presented near the end of the 19th century and has not stopped
since then. Pierce (1905, 2000) and James (1975) describe pragmatism as a whole philosophy
that includes a particular understanding of truth. Truth is different from reality in that its
significance is practical. Reality exists but its structures and relationships are ontologically
distinct from our understanding of the world. In other words, there is an external world that
we do not fully perceive (Cherryholmes, 1992). On the contrary, the legitimacy of a particular
depends on its practical implications and significance (Cherryholmes, 1992). As
Cherryholmes explains, pragmatism shifts “attention to the importance of the consequences of
action based upon particular conceptions” (1992, p. 13). Pragmatism stresses the importance
of reflecting on the practical consequences of scientific activities because scientific inquiry is
likely to influence political decisions and social dynamics.
In this work, we embrace the pragmatist tenet that truth matters to the extent that it concretely
changes our direct environment and practices. Then, pragmatist principles match some
personal values about the relevance of scientific inquiry. As a result, our rationale for
choosing the topic of the dissertation follows a pragmatist conception of scientific inquiry.
We chose to study ICT support to crisis response because we had assessed the need for more
practical guidelines in crisis management and we wanted the dissertation to be potentially
helpful to managers in critical situations.
Therefore, we acknowledge the need for special attention to the relevance of our work
(Lyytinen, 1999a). While Lee (1999) assesses relevance on the degree of the audience’s
interest in the topic of the research, Lyytinen (1999a) highlights the need to meet managerial
needs. Benbasat and Zmud (1999) go further and argue that the relevance of research partly
depends on the ease of its implementation. However, relevance has been criticized for being a
seducing but unreachable objective (Kock et al., 2002). For this reason, one of our goals in
this work our work is to formulate reliable and clear managerial guidelines from our research
findings.

  88
 Section 2. …that calls for a realist ontology

We rely on Blaikie’s classification of different philosophical streams that compose realism

More specifically, our posture corresponds to depth realism, summarized by Blaikie in the
following statement (Blaikie, 2007, p. 13):
“In the depth realist ontology, reality is seen to consist of three levels or domains: the
empirical, the actual and the real (…) Seeing reality of these three domains, ranging from
what can be observed to an underlying domain of causal structures and mechanisms, suggests
the idea of ontological depth, of the stratification of reality that is independent of our
knowledge of it (…). The aim of a science based on this ontology is to explain observable
phenomena with reference to underlying structures and mechanisms (…) Social reality is
viewed either as social arrangements that are the products of the material but unobservable
structures of relations”.
We rely on Outwhaite’s description of the realist ontology principles (2003, p. 282):
“1. The distinction between transitive and intransitive objects of science: between our
concepts, models etc, the real entities, relations and so forth which make up the natural and
the social world.
2. [There is a stratification] of reality into the domains of the real, the actual and the
empirical (…)
3. The conception of causal relationships as tendencies, grounded in the interactions of
generative mechanisms.
4. A rejection of both empiricism and conventionalism. The practical expression of realism is
the concept of real definition. Real definitions are neither summaries of existing verbal usage
nor stipulations that we should use a term in a particular way.
5. The realist concept of explanation involves the postulation of explanatory mechanisms and
the attempt to demonstrate their existence”.
Our ontological assumptions are consistent with Outwhaite’s description of realism. First, we
assume that reality does not only take its essence in our perception but also preexists
independently to our cognition. However, as Outwhaite explains (2003, p. 283), “the social
situations do not exist independently of the way they are interpreted by those involved in
them”. In other words, reality is intransitive in that it has an independent existence from
human subjectivity. Reality is also transitive, which means that researchers have no choice
but to observe it indirectly through the social world.
Crisis is an intransitive object in that its occurrence does not depend on human subjectivity. In
addition crisis can be objectively evaluated in terms of costs and damages. Despite crisis
being intransitive, researchers observe critical situations through the organizational dynamics
they cause, which makes of crisis a transitive phenomenon.

  89
Similarly, improvisation can be defined in objective terms: a short delay between a decision
and an action. Even though objective measurements of organizational improvisation exist,
observing the process of improvising requires in depth analysis that involves the researcher’s
experience most of the time. For instance, researchers have primarily observed improvisation
through artistic practices such as music and theatre. In this sense, improvisation is both
intransitive and transitive.
According to our ontological assumptions, realists researchers deal with three distinct
ontological components: the “real”, the “actual” and the “empirical” (see ontological realist
principle n°2, p. 37). As we previously explained, reality is distinct from what individuals
perceive or understand. The “real” is intransitive while human common sense is the “actual”.
The empirical refers to individual experience. The empirical and the actual are transitive
components of our reality (see ontological realist principle n°1, p. 37). In contrast to the
actual and the empirical, the real is intransitive and is not directly observable (see ontological
realist principle n°1, p. 37)
Our being depth realist has some implications on the type of knowledge we intend to develop
in our work. In spite of the real being not directly observable, realism attempts to understand
the “real” by identifying how the real is generated (see ontological realist principle n°3, p.37).
For this reason, the realist approach excludes pure empiricism (see ontological realist
principle n°4, p. 37). Similarly, realism rejects conventionalism because the aim of realist
researchers is to approach, even in an approximate fashion, the essence of the functioning of
the real rather than producing knowledge that would be unrelated from the real (see
ontological realist principle n°4, p. 37). Even tough the real can be understood in multiple
ways, realist researchers’ choice for an interpretation of the real depends on the practical
implications of this interpretation. As Outhwaite (2003) explains, “Any assertion about [the
real] is good as any other” (see ontological realist principle n°5; p. 37). In other words,
scientific realism is consistent with a pragmatist view of research.
Realism is usually associated with critical realism but it is important to note that our posture
is not critical. The critical dimension of theory implies the existence of previous knowledge
that can be criticized and confronted within a knowledge building process (Edgley, 2003).
Therefore, we agree with Edgley (2003) that the role of science is not necessarily to question
ideologies and conceptions in the society. In our view Science should not substitute for
philosophy6.

Section 3. Some epistemological implications

First of all, realists have to pay particular attention to their biases. This is way we focused our
attention on our own biases during the empirical investigation (see chapter 3.1).

6
We acknowledge there is a debate about the role of critics in research. We agree with Daniel Robey
that critics are necessary to evaluate findings knowledge claims. However, given that knowledge
building is our priority, we restrict our criticism to the issues of validity and coherence of academic
work.
  90
Secondly, one of the fundamental principles underlying pragmatism is the clarity of thought
(Hookway, 2004) that implies the need to provide an integrative and complete explanation of
ICT support to improvisation in crisis. To do so, we connect our observation to existing
knowledge on the same phenomenon so as to provide the most complete account of ICT
support to crisis response. The principle of clarity of thought therefore led to an extended
review of literature in the theoretical part of this document. As James explains (James, 1897,
1975), to be clear, one’s thought should not tend towards oversimplification but rather “class
the diverse cases of a common essence which you discover in items” (James, 1897, p. 67). To
that extent, two main steps are easy to identify in our work: First of all, we lay an extensive
but not exhaustive conceptual basis. Then, we lead an empirical investigation to confront our
theoretical findings and extend our reflection in a theory building approach.

Conclusion

We follow pragmatist and realist principles in this work. We aim to attain reality as an
intransitive object through human actions and perceptions. In addition the pragmatist
approach compels us to account for the practical implications of our findings. For this reason,
we describe the managerial implications of our findings that we submitted to managers in the
final step of our analysis.

  91
  92
 Part 2. The Theoretical assumptions and proposal for
theory
The theoretical part reports some preliminary work on the concepts of crisis response,
improvisation and communication technologies (DeSanctis, Poole). The objectives of this part
are four-fold.
- Set the definition of the concepts we use in this work, such as crisis, crisis response
and organizational improvisation,
- Analyze existing literature on crisis and improvisation and detail the basic theoretical
assumptions we extract from it,
- Argue the choices and directions we take in our work (presentations of our unit of
analysis and research questions),
- Make some theoretical propositions to respond to research question I.
This part is therefore structured on the following path:
In chapter 2.1, we present our basic assumptions on crisis and justify our focus on crisis
response in large-scale disasters. By doing so, we narrow the scope of our investigation to a
specific type of crisis context.
In chapter 2.2, we tackle the notion of improvisation by a literature review and classification.
Improvisation is a broad concept that gathers different themes and phenomena. On the basis
of a literature review, we identify eight coexisting perspectives on improvisation in
management research.
In chapter 2.3, we propose some definition of organizational improvisation that brings into
the open its characteristics in terms of interindividual interactions. This theoretical proposal is
rooted in the dynamic capabilities approach. For the basis of our definition, we review the
literature on the ICT contribution to organizational improvisation in crisis response. Our
findings from this review reveal five technological characteristics that potentially bolster
interactions that comprise organizational improvisation. Our discussion suggests the need for
a concrete observation of interactions and ICT use during crisis improvisation, which is
presented in the third part of the document.

  93
Chapter 2.1. The context of investigation: Presentation of the scope of
enquiry and theoretical assumptions on the concepts of crisis and
crisis response

Introduction

This chapter presents our assumptions in relation to crisis. Various conceptualizations of

crisis coexist in the management literature (Pauchant, Douville, 1993). For this reason,
tackling the concept of crisis has led us to make the following assumptions.
First of all, we approach crisis as a specific organizational episode articulated by four steps:
crisis latent development, immediate response, crisis coping and crisis recovery.
Secondly, we focus on the second and the third steps in the organizational crisis cycle that are
immediate response and crisis coping. In other words, we address the response that follows
the triggering event and aims at mitigating crisis immediate consequences.
Thirdly, we narrow our analysis to crisis response that happens in the context of natural
disasters provoked by natural hazards.
This chapter is structured as follows. Section 1 presents a review of the crisis management
literature and accounts for major theoretical divergences. Section 2 details some key
characteristics of crisis and our theoretical positioning on the definition of crisis as an
organizational cyclic episode. We then explain our focus on crisis response in section 3. We
argue our intention to study crisis response in the context of natural disasters provoked by
natural hazards in section 4.

Section 1. Crisis research as a developed but heterogeneous field

Crisis is a generic concept. There is a long list of disparate events and situations that one can
consider as critical and estimate as valuable for research (Roux-Dufort, 2007). From industrial
catastrophes such as Bhopal (Shrivastava, 1987), or the Guadalupe Dunes spills in California
(Beamish, 2002), to terrorist attacks (Mendonça, 2007), to organizational and technical
accidents (Hutchins, 1991; Perrow, 1999; Weick, Sutcliffe, 2001), to firefighting failure
(Weick, 1993) and disease breakouts (Leidner et al., 2009). However, it is necessary to clarify
the aspects of crisis that we study.
Crisis is generally defined as a critical and stressful condition that threatens major goals of an
organization (Hermann, 1963; Perrow, 1984; Weick, 2001a) and that can be triggered by
human or technological errors (Pearson, Claire, 1998).
First, each crisis is unique and it is difficult to find common features to very different
situations, ranging from financial breakdowns to natural disasters (Pearson, Claire, 1998;
Roux-Dufort, Vidaillet, 2003b). Table 5 illustrates the wide array of events that can trigger
crisis. One can deduce from this list of events that one crisis can have multiple causes. For
example, during 2005 Katrina, the disruption of services and products simultaneously
  94
happened with the destruction of the organizational information base, escape of hazardous
materials, personnel assault and malicious rumor.

Table 5. Organizational crises and triggering events as defined by Pearson and Claire
(1998)

Crisis research involves many divergent perspectives (Boin, 2004). Given the wide diversity
of crises, researchers have discussed the very definition of crisis. In spite of this effort, deep
divisions remain within the research community (Roux-Dufort, 2007).
First of all, there is a debate on whether crisis is endogenous or exogenous. On one hand,
crisis can be seen as an extraordinary event that unexpectedly endangers an organization’s
survival (Hermann, 1963; Pearson, Claire, 1998; Pearson et al., 1997). On the other hand,
crisis reveals the inability of organizations to adapt to an expected situation because of
internal structural deficiencies or disequilibrium (Mitroff et al., 1989; Morin, 1976; Probst,
Raisch, 2005; Roux-Dufort, 1997; Weick, Sutcliffe, 2001). Even in threatening situations,
some organizations succeed in responding. This shows that “being in crisis” is not only
related to a specific event but also tied up to the organization ability to cope either with its
environment, its fundamental objectives (Fink et al., 1971) or its “harmful activities”
(Shrivastava et al., 1988b).
There is another debate on the distinction between regular difficulties that occur in
organizational life and crisis. Crisis is specific in that it threatens an organization’s
fundamental values and survival (Hermann, 1963; Nyström, Starbuck, 1984). It is undeniable
that experiencing major difficulties is a part of business life. However, researchers distinguish
  95
crisis from casual difficulties by its stakes, such as organizational survival and social cost.
They define crisis as a situation where “many members of a social system fail to receive
expected conditions of life from the system” (Barton, 1969, cited by Quarantelli, Dynes,
1977). For example, bankruptcy is part of economic life but becomes a crisis since it burdens
social groups.
Similarly, researchers’ opinions diverge on crisis existence being objective and subjective.
Some authors systematically refer to crisis as an objective reality (Pearson, Claire, 1998;
Probst, Raisch, 2005) and define it on the basis of quantifiable criteria such as economic cost
and the number of fatalities. Figure 20 represents divergences among authors in the crisis
management literature.
However, we think crisis existence is both objective and subjective (Billings et al., 1980;
Milburn et al., 1983a; Roux-Dufort, 2007). Despite objective measures, crisis is also a
personal experience (Milburn et al., 1983b; Mitroff et al., 1996; Starbuck et al., 1978). More
specifically, we define crisis as an organizational episode in which an organization is unable
to respond to an unexpected situation that threatens its functioning and is likely to have
negative social outcomes. We further detail the elements of this definition in the following
section.

  96
 Figure 20. Divergence among crisis researchers

Section 2. Our basic assumptions on crisis

2.I. Crisis: an organizational phenomenon

We study crises from the inner perspective of organizations. To do so, we examine
organizational activities that are specific to a critical situation (Fink et al., 1971; Richardson,
1994; Shrivastava, 1993). In addition, we adopt a systemic approach, inspired from Katz and
Kahn’s definition of organization in The Social Psychology of Organizations (1966). This
definition depicts an organization as an open and complex system of interdependent
subsystems. These subsystems interact with each other to meet common objectives.
According to this view, crisis corresponds to an inner disequilibrium between resources that
the whole organizational system has to cope with (Morin, 1976).

  97
 2.II.Social and natural Costs
We assume that crisis directly or indirectly induces social costs (Barton, 1969). For instance,
an organizational failure can be considered as a crisis if it directly or indirectly implies higher
costs than lower performance, such as material damage, increased unemployment or human
loss. For instance, the 2003 French Heat Wave caused 15,000 fatalities (Lalande et al., 2003);
the Guadalupe spills spread 20,000,000 gallons of oil in the Pacific Oceans jeopardizing the
health of millions of people (Beamish, 2002). In 1984, the Bhopal disaster killed over 3000
people and injured another 200,000 (Shrivastava, 1987). The explosion of the Challenger
Shuttle provoked the death of 8 people but also disrupted the activity of the NASA and
provoked psychological difficulties among the victims’ colleagues (Mitroff et al., 1996).
2.III. Surprise
Crisis differs from routine in terms of objective and subjective specificities. Surprise is one of
these subjective specificities and the second key characteristic of crisis (Hermann, 1963).
Surprise refers to the fact that a crisis, by definition, is hardly anticipated. Part of the surprise
comes from the triggering event, the third key characteristic of a crisis situation, that provokes
the crisis situation (Shrivastava, 1994). The triggering event is a seemingly insignificant and
low-probability event (Pearson, Claire, 1998) that makes anticipating crises even more
difficult. Given organizational flaws, the triggering event’s impact is strong (Pearson, Claire,
1998) and opens into a crisis situation that also surprises managers. Bhopal: Anatomy of a
crisis (Shrivastava, 1987), narrates the very beginning of the Bhopal Disaster.
“At the core of any industrial crisis, is a triggering event. In Bhopal, the triggering event was
the leakage of a toxic gas, MIC, from storage tanks. Human, organizational, and
technological failure in the plant paved the way for the crisis that ensued” (p.35).
Surprise differentiates crisis from emergency situations. Emergency is characterized by a need
for a timely action (Mendonça et al., 2004a). However, surprise in crisis implies new
constraints for action and less room to follow routine, which does not necessarily characterize
all emergency situations.
2.IV. Emotional and time pressure
Crisis is characterized by high emotional and time pressure that requires an innovative
response (Billings et al., 1980; Milburn et al., 1983b; Smart, Vertinsky, 1977; Staw et al.,
1981).
Crisis responders’ emotional pressure comes from their experience of damages and human
suffering, as well as the lack of adequate resources to cope with the situation, such as
material, information or time (Starbuck et al., 1978). Emotional pressure is highly subjective
and is expressed by stress or even panic (Hall, 1971; Milburn et al., 1983b).
People have various reactions to emotional pressure. Some authors explain that stress can
imply denial of the situation (Hart, 1993; Perry, 1991b), threat-rigidity (Smart, Vertinsky,
1977) and inconsistent decision making (Starbuck et al., 1978). Emotional pressure can
positively influence crisis response as well by making individuals hyper vigilant (Milburn et

  98
al., 1983b) or incredibly creative (Ciborra, 1996b; Weick, 1993). During the Mann Gulch
Disaster, one of the smoke jumpers spontaneously had the idea to burn a circle for self-
protection. Claudio Ciborra, in the Labyrinths of Information (1996b), highlighted the
ingenuity of this initiative and explained that survival instinct led the smoke jumper to come
up with a solution. Emotional pressure is a factor of increased creativity and the ability to
improvise during crisis.
Time constraints come from the fact that time is a scarce resource during crisis (Ahituv et al.,
1998). The influence of time pressure on information processing was evidenced (Beach,
Mitchell, 1978; Galbraith, 1974; Rieskamp, Hoffrage, 2008; Weenig, Maarleveld, 2002). One
could safely assume that crisis responders make bad decisions when they lack time. However,
by making individuals more vigilant, time constraints can have a positive effect on the quality
of information processing (Maule, Edland, 1997).
To conclude, emotional pressure and time constraints are key characteristics of crisis response
and have mixed effects on crisis responders. Managed properly, emotional and time pressure
may help organizations to be more reactive to triggering events, such as in high-reliability
organizations (HRO), (Weick, Sutcliffe, 2001).
2.V. An objective and subjective existence
Crisis is inherently subjective in that it depends on one’s ontology, risk aversion and personal
life (Billings et al., 1980; Pauchant, Mitroff, 1992). Depending on their perception of the
situation as a crisis or not, people may tend to follow routine activities or develop more
resilience (Roux-Dufort, Vidaillet, 2003b). Similarly Mullen, Vladi and Mills (2006) argue
that the personal dimension of crisis is important. Pauchant and Mitroff (1992) expand on that
argument by explaining that the Challenger staff experienced a “psychological death”, which
has a major impact on crisis response. In the psychology literature, Caplan (1964) defines
crisis as “what happens when a person faces a difficulty in which coping repertoire is
insufficient”.
Crisis comes not only from self-perception but also from the perception of one’s work
practices (Roux-Dufort, Vidaillet, 2003b). People may acknowledge the situation as more
critical because they have to achieve new tasks, or because routine activities are not relevant
to the situation (Crossan et al., 1996), or they have to coordinate with counterparts they did
not know before (Perry, Quarantelli, 2005), or even because regular activities have to be
completed in hurry. People do not always perceive the triggering event as the beginning of a
critical situation and most of the time the process of making sense out of crisis is retrospective
(Weick, 1988). In other words, a crisis situation may be more easily recognized by the shift of
organizational functioning. This shift can be fragmented in four periods that characterize
crisis from the organizational perspective. For instance, individuals more easily perceive
crisis once crisis response and recovery takes place (Turner, 1976). Crisis response and
coping corresponds to two stages of the development of crisis (Pearson, Claire, 1998;
Shrivastava, 1993; Turner, 1976). The next section presents this model.

  99
 2.VI. Organizational crisis as a cycle
Table 6 presents the specificities of the different stages of the development of an
organizational crisis and specific issues for each stage. Before crisis concretely manifests,
latent dysfunctions appear and existing precautions may be violated (stage 1) (Turner, 1976).
For example, before the merger between Union Pacific (UP) and South Pacific (SP) in August
1996, expertise was declined and minor accidents occurred (Weick, Sutcliffe, 2001). When
negative signals are not detected, crisis is less likely to be avoided (Beamish, 2002;
Shrivastava, 1988). Once the triggering event occurs, the organization reacts to the situation
without making sense of what is happening (stage 2). Emotional pressure, as described herein,
intensifies the personal experience of organizational crisis. The coping stage is characterized
by the emergence of organizational processes that aim at limiting damage (stage 3). The
organization gets “organized” to cope with the situation (Weick, 2001c). Finally, stage 4
gathers crisis recovery, organizational change and post-crisis learning. This stage is not taken
into account in our work.

  100
 Table 6. Organizational crisis stages and corresponding issues

Organizational crisis Organizational issues definition

stages

Latent crisis - Inner dysfunctions are likely to provoke decline (Cameron et

al., 1987) and they interact with each other (Perrow, 1984).
- Agency issues and information retention occur (Beamish,
2002).
- There is a lack of vigilance (Weick, Sutcliffe, 2001) and
unexpected accidents occur (Beamish, 2002; Roberts, Bea,
2001; Weick, 1988).
- Signal detection management is a crucial determinant of crisis
response (Shrivastava, 1988).

Immediate response to - Emotional shock, psychological experience of death and

triggering event breakdown of previously developed representations (Mitroff et
al., 1996).
- Possible denying of crisis (Fink et al., 1971) and rigidity threats
(Staw et al., 1981).
- Collective perception of difficulties to cope with high
uncertainty and complexity (Billings et al., 1980; Milburn et al.,
1983a).
- Risk for organizational instability (Pauchant, Mitroff, 1992).

Coping and response - Divergences on how to organize crisis response: the Importance
period of centralization (Hermann, 1963; Staw et al., 1981), loose
coupling (Robert, Lajtha, 2002) or structural redundancy
(Roberts, 1990a) and introduction of the crisis facilitator.
- Creation of emergency management teams or crisis decision
units (Smart, Vertinsky, 1977). Additional subsystem to treat
uncertainty and crisis events (Katz, Kahn, 1966).
- Actions are taken in hopes that one will just happen to reduce
the crisis (Starbuck, Greve, Hedberg, 1978, cited by Fiol, Lyles,
1985).
- New informational processes (Osborn, 1998) and growing
informational needs (Smart, Vertinsky, 1984).
- Possible emergence of boundary spanning activities (Leifer,
Huber, 1977).
- Preplanned crisis procedures may not be relevant to the
situation (Pauchant, Mitroff, 1988) or not used (Perry, 1991b)
and improvisation can occur.
- There is an organizational adaptation (Fink et al., 1971),

  101
 responsibilities are renewed depending on political conflicts
between stakeholders (Leifer, Huber, 1977).
- Interruptions in the decision making process (Hale et al., 2006)

Post Crisis - Organizational learning or unlearning from crisis (Fauchart,

2006; Stead, Smallman, 1999).
- Positive or negative outcome (Clair, Dufresne, 2005; Seeger et
al., 2005).

Section 3. Crisis response: some theoretical background

3.I. Crisis response: a paradoxical process

The word “crisis” derives from the Greek word krisis that means “decision” (Morin, 1993;
Pauchant, Douville, 1993). This implies the importance of managers’ understanding of crisis
and its outcomes. Research on crisis calls for further developments related to crisis response
(Schoenberg, 2005). Part of the research has been more focused on crisis preparation than
crisis response so far. During crisis response, planned procedures should be applied (Pearson
et al., 1997) but also adjusted as necessary because every crisis is unique and prevention can
never be completely adequate for the situation (Quarantelli, 1988; Waugh, Streib, 2006).
At first glance, crisis response may seem tautological because it simultaneously characterizes
and aims at solving an ill-structured situation; otherwise it would not be crisis response.
However, we define crisis response as the set of decisions and actions that lead to making
sense of the situation and possibly avoiding too much resulting damage. We disagree with
Mishra (1996) as we think crisis response cannot be considered routine. Contrary to routine,
crisis confronts organizations with events they are not prepared for and questions their ability
to keep things under control. For example, the 2003 French heat wave questioned medical
personnel competencies and ability to save lives. Because physicians and nurses lacked
knowledge on hyperthermia, they were unable to diagnose and treat patients effectively for
the first time in years. Furthermore, organizations can have to operate outside of their routine
standards when crisis responders are numerous. For instance, organizations have to quickly
relate to unfamiliar entities in disasters. Additionally, some new activities are required but can
reduce the pace of response (Quarantelli, 2006). For example, during the French Heat Wave,
hospital managers had to invent new procedures when they searched for extra spaces to place
corpses.
Finally, people are likely to get confused during crisis response due to the ambiguity of the
causes, effects and means of resolution of the crisis (Pearson, Claire, 1998):
“A paradox is associated with crisis management: We cannot fully understand what we need
to do unless we first understand what need to do and have in place before a crisis. At the
same time, we cannot understand fully what we need to do beforehand unless we understand
what we will be required to do during a crisis” (Mitroff et al., 1996, p. 59).

  102
Despite the need to act promptly (Pearson, Claire, 1998), crisis responders are uncertain of the
outcomes of their decisions (Fink et al., 1971; Nyström, Starbuck, 1984). More generally,
crisis challenges individual ability to make sense of events (Weick, 1988).
3.II.Improvisation in crisis response
In the literature, improvisation is somewhat associated with crisis response. However, the
conceptual ties between these two concepts are vague. Both conditions of emergence and
effects of improvisation in crisis response remain uncertain.
Improvisation logically emerges from time pressure (Crossan et al., 2005), the lack of
adequate planning (Rerup, 2001; Suchman, 1993) or a need for the continuous adjustment of
action (Crossan et al., 2005; Mendonça, 2007). Improvisation is also likely to happen because
the top management’s legitimacy is questioned, which endorses the right to alternative paths
of action (Benson, 1977). However, the probability of improvisation is not linearly correlated
with uncertainty (Cunha et al., 2003).
Good crisis response requires resilience, which requires improvisation (Rerup, 2001). Crisis
responders have to quickly adapt to unexpected needs in terms of leadership, decision making
and coordination. In addition, every crisis requires a specific response that cannot be always
prepared (Waugh, Streib, 2006). Therefore, improvisation brings about innovation that is
necessary for crisis response (Fink et al., 1971). For instance, during the 9/11 response, the
decision to connect electric cables of telecommunications infrastructures to diesel generators
“took milliseconds” (Mendonça, 2007) but deeply enriched the actors’ repertoires of action.
Finally, improvisation contributes to the renewal of routines (Feldman, Pentland, 2003).
However, if not managed properly, improvisation can burden coordination and therefore
endanger human lives by bringing more complexity and ambiguity in crisis response (Perry,
Quarantelli, 2005). Furthermore, an aborted improvisation boils down to a waste of time for
crisis responders. Academics and managers remain uncertain that learning from past
improvisation is feasible, due to its extemporaneity. As Ryle (1979) describes: “What comes
to pass on one occasion has, with all its concomitants, origins and details, never taken place
before and will never take place again”.

Section 4. Crisis response in the context of natural disasters

Disasters and crises are distinct phenomena (Shaluf et al., 2003). As disasters are increasingly
frequent (Perrow, 2006; Torry, 1979), they represent a considerable proportion of crisis
response nowadays. The economic cost associated with natural disasters has increased 14-fold
since the 1950’s (Guha-Sapir et al., 2004). For this reason, we focus on crisis response in the
context of natural disasters that are provoked by natural hazards (Torry, 1979).
The conception of disaster was originally generic (Quarantelli, Dynes, 1977), but disasters
differ on the basis of their cause and their outcomes (Gephart, 1984). There are divergent
views on the connections between disasters and crisis (Shaluf et al., 2003). For some authors,
disasters are the direct consequences of some organizational crisis that was not properly
handled (Lin et al., 2006). For others, disaster is a context in which organizational crisis can
  103
happen (Pearson, Claire, 1998). We follow the latter conception on disasters. A disaster is
characterized by its physical consequences, such as human loss and material damage, and the
social disruption it provokes (Comfort, 1993; Dynes, 1970).
The onset of the natural disaster (also called the physical agent by Quarantelli and Dynes in
1977) can be at the origin of the organizational crisis. Disasters destroy infrastructures,
disrupt services, damage organizational information bases and sometimes eliminate key
stakeholders (Pearson, Claire, 1998). Ives and Junglas (2006b) provide a good example on
how disasters unavoidably provoke organizational crisis. They depict how the Northrop
Grumman Corporation, which employed more than 20,000 in New Orleans in August 2005,
had to manage the aftermath Katrina left behind. Employees were missing and a
communication blackout occurred. The corporation intended to help its workers, which led
the IT director to go to New Orleans to search for and evacuate employees and organize the
IS recovery. Organizations are likely to be brutally affected by disaster, which puts them into
a crisis because they need to cope with unanticipated damage or loss of resource.
Furthermore, organizations have to cope with numerous but unfamiliar groups and individuals
who are affected by the disaster (Shrivastava, 1988).
Crisis response in the context of natural disaster covers two main steps that correspond to two
particular periods in the organizational crisis cycle: the immediate response to triggering
events and coping with crisis. Crisis response in natural disaster is therefore likely to be
complex. Also, disasters induce “breaks in pattern” and “isolating interruption of norms”
(Wallace, 1956, cited by Torry, 1979). In that sense, improvisation, or at least strong
deviation from established patterns of action is likely to happen during crisis response in
disaster settings.

Conclusion

In this chapter, we present the context of our investigation. We approach crisis through the
concept of organizational crisis. We restrict our investigation of organizational crisis to the
stages of crisis response and crisis coping. Finally, we focus our interest on disaster situations,
in which interactions are crucial but complex and unpredictable.
Table 7 presents the key characteristics of the type of crisis we intend to study: organizational
crisis response in the context of natural disasters provoked by natural hazards. Each of these
characteristics stands for basic assumption in our work.

  104
 Table 7. Key characteristics of disasters, crises and crisis response

Source of characteristics Characteristics of the situation under study

in the literature

Natural disaster provoked The triggering event (or the physical agent) is natural and partly
by natural hazard related to natural hazard.

Disaster An important diversity of organizations and actors are involved

in the response

Disaster Actors have to quickly develop coordination with unfamiliar

organizations

Crisis response Emotional pressure is high and likely to impact the coping
process

Crisis response Crisis requires innovative response from organizations. By

definition, routine and previously defined procedures cannot be
blindly followed

Organizational crisis cycle Crisis life can be divided in four periods: latent dysfunctions,
triggering events, coping period and post crisis period.7

7
We assume in our work that information technology failure is not the main factor in the triggering event of the

crisis or the technological error. By doing so, we still consider information technologies’ potential support to the

organizational response of crisis more easily.

  105
Chapter 2.2 Coping with diversity in research on improvisation8

Introduction

The concept of improvisation is the focus of an increasing number of researchers. First, as a

daily practice, improvisation deeply influences organizational life (Barley, 1986; Brown,
Duguid, 1991; Ciborra, 1996b). Second, improvisation is a relevant concept to organizations
that need to be resilient when faced with increasing environmental uncertainty and time
constraint, (Weick, 1998). Moreover improvisation challenges researchers (Ciborra,
Willcocks, 2006) in that it opens a new window to themes such as “ambidexterity”,
“tinkering”, “bricolage” and “organizational paradoxes” (Cunha, 2001; Vera, Crossan, 2007).
For these reasons, it is necessary to better understand it.
The widespread use of the concept of improvisation emphasizes its relevance in multiple
disciplines, ranging from strategic development (Mintzberg, McHugh, 1985), entrepreneurial
projects (Baker, Nelson, 2005; Hmieleski, Corbett, 2008; Jiyao, Jing, 2005), product
development (Akgün, Lynn, 2002; Cunha, Gomes, 2003; Moorman, Miner, 1998b), to crisis
response (Hutchins, 1991; Lanzara, 1983; Mendonça, Wallace, 2007; Weick, 1993), or even
Information system development (ISD) and ICT use (Hanseth et al., 1996; McGann, 2005;
Pawlowski et al., 2004). Undoubtedly, a significant amount of work has settled some
commonly shared definitions of improvisation. For instance, Cunha, Cunha and Kamoche
have provided an integrative definition of improvisation on the basis of an exhaustive
literature review (1999c). As a result, authors commonly agree that improvisation consists of
an innovative use of resources in a restricted delay between decision and action.
However, research on improvisation seems varied. Researchers have used the concept of
improvisation in numerous ways, targeting separate scopes, contexts and motives for enquiry.
For instance, some of them use improvisation to label unexpected practices, such as Faia-
Correia (2003) whereas others present it as a new paradigm (Hatch, 1999). A more recent
stream of research explores the cognitive processes that enable the emergence and the
development of improvisation or explore it from a cognitive standpoint (Mendonça, Wallace,
2007).
Even though diversity fosters creativity, researchers can experience difficulties in making
sense of the variety of academic work on improvisation. Studies on improvisation are not only
diverse from a theoretical perspective, but also include distinct rhetorical and metaphorical
techniques. As an example, Bigley and Roberts’ work (2001) has little if nothing to do with
Pava’s use of the concept of improvisation (2002). While the former authors explore the
impact of Incident Command System (ICT) on improvisation in emergency situations, the

8
This study was completed in collaboration with two other authors. Dr. François-Xavier de Vaujany
participated in the design of the classification scheme, the cross coding and the choice for statistical
methods. Dr. François-Xavier de Vaujany et Dr. Christophe Elie-dit-Cosaque participated in the
formulation of research questions, the refinement of the classification scheme and the structure of the
document.
  106
latter proposes the original concept of ethical improvisation based on an interpretation of the
Bible. How then can researchers compare findings without evidence that the studies in
question relate to the same aspects of improvisation?
Diversity can transform heterogeneity, therefore preventing discussion between authors to
challenge each other’s view of improvisation. As a result, researchers have missed discussion
on crucial aspects of improvisation, such as the ins and outs of the coordination that takes
place between improvisers. As building relevant knowledge on improvisation has become
progressively more challenging there is a need for an overall picture of the similarities and
dissimilarities between studies on improvisation. This work aims to respond to this need by
addressing the following questions:
- Is research on improvisation effectively diverse?
- How to account for diversity in research on improvisation?
We addressing these two research questions by reviewing and classifying the management
literature on improvisation. Literature reviews are relevant beginnings to tackle emergent
issues in academic life (Webster, Watson, 2002). One could argue that literature reviews on
the concept of improvisation are not new. Moorman and Miner (1998a) led a first literature
review on improvisation. They categorized the different fields in which improvisation was
analyzed. Cunha, Cunha and Kamoche (1999b) pursued this work by reviewing existing
definitions of improvisation. Since then, literature on improvisation has been intensively
developing.
Our work differs from previous literature reviews on improvisation in that we use quantitative
techniques to provide an updated and revised snapshot of the literature on improvisation. We
rely on the classification principles defined by Bowker and Starr (2000) to categorize primary
aspects of the literature on improvisation. We then use CATPCA, PCA and cluster analysis
techniques on a sample composed of 105 articles. Our analysis suggests two types of findings.
First of all, we could not identify any predominant paradigm in research on improvisation. In
addition, conceptualizations of improvisation are relatively similar across various periods of
time. In our view, these findings mean that research on improvisation is effectively diverse
and that no one approach to improvisation has become predominant over time. Secondly, we
suggest that diversity originates from divergences between authors in relation to four primary
research tasks. First of all, authors rely on different techniques to evidence improvisation
conceptually and empirically. While some of them deduce improvisation from existing
frameworks, others induce their own understanding of improvisation from data. Secondly,
authors do not investigate the same kinds of improvisers nor the same practices related to
improvisation. Some authors focus on top management improvisation in designing. Other
authors deal with improvisation practices by individuals from various hierarchical layers.
Thirdly, authors contextualize improvisation differently. Some of them argue that
improvisation endogenously emerge in more or less turbulent environments. However, some
authors have depicted improvisation as a spontaneous response to crisis in the last years.
Finally, the authors pursue divergent objectives in terms of knowledge creation. Some of

  107
them aim to contribute to a unified representation of improvisation while some others offer
alternate views. This work is structured as follows. Section 1 discusses the importance of
addressing diversity in management research and presents the aspects of diversity we consider
in this study. We then detail the analysis process and the statistical techniques we use in this
work in Section 2. Finally, Section 3 details our interpretation of the findings, followed some
discussion.

Section 1. Theoretical background

1.I. Diversity and knowledge building

Diversity has been widely documented in the management literature. In particular, Argyris
and Schon (1974) explain how individuals appropriate concepts by developing their own
theories, therefore contributing to divergent views of the same topic. Reflection about the
benefits and costs of diversity within academic life has been pursued (Pfeffer, 1993), at least
since the publication of Koontz’s “The management theory jungle” in 1961 (Mathews et al.,
1999). In the MIS field specifically, diversity has been questioned. While some authors argue
that addressing a wide range of perspectives promotes relevant research and progress,
diversity can also hamper the identity of a discipline (Benbasat, Weber, 1996).
Scholars tackled the concept of diversity from either a descriptive or a normative standpoint.
From a descriptive standpoint, a significant effort has been led to evaluate diversity. For
example, Bakhri and Sheetz evaluate the frequency of use of major theories in the MIS field
(2001). Tsai and Wu also contribute to this effort by exploring the effect of co-citation on
diversity in the management field (2010). Additionally, some authors focus their exploration
of diversity on specific topics. Numerous examples such as Schultze and Orlikowski’s work
on the concept of virtuality (2001) or Cunha and his colleagues’ work on improvisation
(1999c), illustrate the ongoing effort to define and evaluate the diversity of studies on specific
management topics. Diversity is also controversial as some authors make divergent claims
about academic work standards. For example, Benbasat and Zmud called for an explicit
discussion about whether academics should focus on a unique set of problems, theories or
methods in the MIS disciplines (Benbasat, Weber, 1996). On the contrary Bamberger and
Pratt (2010) advocate for diversity and emphasize the need to promote unconventional
studies. Some other authors provide more practical recommendations, such as Robey (1996)
who argues that methodological rigor is necessary to promote diversity.
The underlying question of this debate is the effect of diversity on knowledge building.
Diversity of perspectives in regards to specific concepts nourishes conceptual debate and is
inherent in scientific progress. As Khun explains (1962), scientific revolutions stem from the
temporary challenging of existing paradigms by alternate perspectives. Similarly, Popper’s
principle of falsification relies on the tenet that science is not restrictive to one unique
approach (1972). Undoubtedly, approaching reality from a single perspective leaves academic
life with fewer opportunities for debate and understanding. For this reason, diversity is
necessary in research. However, diversity can lead to heterogeneity and cause scholars to miss

  108
opportunities for discussion and the refinement of findings. We propose in the following
paragraphs three solid motivations for providing an integrative representation of diversity
within an academic discipline.
First of all, a common referential is needed for discussion between authors. When studies
have absolutely nothing in common, how does one assess the legitimacy of findings?
Discussion is more difficult to lead when authors offer divergent views without sharing
common concepts or definitions. For instance, a significant amount of work has discussed the
relevance of the theories of ICT fit (Zigurs, Buckland, 1998; Zigurs, Khazanchi, 2008). Even
though these theories provide divergent findings as to the relevance of using technology in
organizations they share a common assumption of fit, e.g. a minimum level consistency
between organizational structures and environment is necessary to attain performance
(Drazin, Van de Ven, 1985). In addition, all theories of fit start with the measurement of
technology features by a distinct construct from task and user characteristics.
Secondly, if not reported, differences within academic work can dissuade authors from
proposing managerial guidelines to improvisers or even contributing to previous managerial
recommendations. Of course, it happens that some authors’ approach has little in common
with some other studies. However, if these authors cannot assess the extent to which their
findings complement or contradict other results, then releasing managerial guidelines is
riskier. Rather than taking the risk of providing contradictory managerial guidelines regarding
improvisation, most researchers do not provide any guidelines at all. For example, Gauthereau
and Hollnagel (2005) recommend that researchers be cautious about managerial guidelines on
improvisation because the very nature of improvisation is unknown. Similarly, Miner and his
colleagues argue that providing managerial recommendations about improvisation is hardly
achievable, due to our lack of knowledge on the specificities of each improvisation (Miner et
al., 2001). As a result, current conceptualizations of improvisation remain unsatisfactory for
managers who lack concrete recommendations such as what resource to leverage for
improvisation. In spite of it’s relevance to the development of personal resilience (Vera,
Crossan, 2004) to face time and organizational constraints (Ciborra, 1999), improvisation is
insufficiently used by managers. In our view, diverse approaches and findings on
improvisation can product consistent managerial guidelines if they are properly documented.
Thirdly, by not being aware of differences between their research, authors tend to overlook
potential sources of insight. An integrative representation of theories can promote synergies
between streams of research and help to resolve complex issues. For instance, most authors
who have addressed improvisation as an organizational phenomenon fall short of details on
how improvisers manage coordination. Interestingly, Weick (1993, 1998) Ciborra (1996a, b)
have brought insights on interactions during improvisation. However, their findings have not
been fully taken into account to clarify coordination dynamics that take place during
improvisation. Similarly, Weick’s reflection on improvisation intensity (1998) has been
overlooked by many studies of improvisation. This work makes no exception: We do not refer
to improvisation intensity in the classification because too few articles mention this dimension
of improvisation. A common representation of the existing approach is therefore necessary to

  109
grant that theoretical achievements be shared or taken into account in the academic field that
deals with improvisation.
1.II.Diversity and research on improvisation
A quick lexicometric search shows improvisation has been deeply examined in a large panel
of disciplines. Simple requests on EBCOHost reveal more than 238 articles include the term
improvisation in their abstracts. A second request on Web of Science reveals 149 articles in
business and behavioral sciences cover improvisation as a topic. This quick investigation also
reveals the variety of works on improvisation. Improvisation practices and cognitive aspects
have been explored in theatre, literature and music, nursing and psychology. In business
research, improvisation is most of the time associated with concepts such as innovation,
learning, product development or technology use. This work focuses on the use of the term
improvisation in business research.
The commonly accepted definition of improvisation refers to an innovative use of resources
and a restricted delay between acting and planning. Some authors characterize improvisation
by it’s innovative use of resources (Weick, 1993, 1998), or by a reduced delay between
decision and action (Ciborra, 1996a; Moorman, Miner, 1998a). This core meaning can be
found in the etymological dictionary: the term improvisation (first identified in 1786) comes
from the French improvisation that means "compose or say extemporaneously", coming from
the Latin improvisus that means unforeseen (in- "not" + provisus "foreseen, provided”).
However, the ontology of improvisation differs from one study to another, coming from a set
of actions (Heeks, 2002a), a will (Miner et al., 2001), a capability (Rerup, 2001)or a process
(Ford, 2008). Furthermore, authors pose different tacit assumptions about improvisation.
While some authors understand improvisation as the expression of individual intuition and
creativity, others pose improvisation as organizational paradoxes (Clegg et al., 2002) and
disruption (Benson, 1977). Finally, studies greatly differ in terms of methods. Some authors
draw inspiration from surprising contexts such as circus, medical structures, jazz bands or
theatre groups to understand improvisation (Meyer et al., 1998). Others restrict their studies
to a strict organizational context (Brown, Duguid, 1991; Faia-Correia, 2003). As a result, it is
unclear as to whether researchers refer to the same concept when studying improvisation.
Undeniably, there is a need to clarify the dimensions of diversity with respect to
improvisation to promote knowledge building. We propose in this work an integrative
classification of research on improvisation. To do so, we preliminary need to answer to the
following question: How does one handle diversity in an academic field?
1.III. Strategies to analyze diversity
In this section we detail our strategy to handle diversity in research on improvisation. We
approach diversity from a realist stance. We analyze authors’ representation of improvisation
to make sense of diversity on this topic.
Generally speaking, academic diversity develops with the coexistence of divergent problems,
theoretical foundations and methods (Barkhi, Sheetz, 2001; Benbasat, Weber, 1996; Koontz,

  110
1980). However, in this work, we adopt a realist stance that posits that diversity is not
restricted to authors’ view of improvisation. Due to the primary assumption that reality is not
directly observable but only partially ceased through human perception, realist researchers
accept that a single term can refer to multiple realities. This means that not only do authors
conceptualize improvisation from multiple perspectives, but improvisation can also refer to
various empirical cases as well. For instance, experts’ improvisation in crisis cells (see
Bigley, Roberts, 2001) and improvisation in top management teams who design new
strategies (see Perry, 1991a) are likely to result into various conceptualizations. Therefore,
handling diversity on improvisation requires examining improvisers and improvisation
contexts. Moreover, authors’ interest for an empirical case of improvisation is dependent on
their ontological and epistemological postures. For example, Clegg, Cunha and Cunha inspire
from Benson’s dialectical view of organizations to conceptualize improvisation as an
organizational paradox (Clegg et al., 2002).
Furthermore, we base our research upon the assumption that the way authors’ find, depict and
analyze empirical cases of improvisation is likely to play a role in academic diversity. To take
into account these aspects of diversity, we analyze the instrumentation of concepts by authors,
e.g. the manner a concept is arranged with others. To develop our point we draw on
Wittgenstein’s principles on semantics, cited by Anderson and Ortony (1975). According to
these authors, one should analyze how a concept is used to understand author’s stance on the
concept as well as its contextual specificities. For instance, they examine how the word eat
conveys various emotions and images, depending on the context of its use. The meaning of
eating something is the same whether its object is a fruit or meat. Meanwhile, in the
imaginary, our representation of eating a fruit or eating meat significantly differs. As
Anderson and Ortony explain:
“Consider the phrases eat steak, eat soup and eat apple. Eating a steak requires a knife and
fork. Soup is sipped with a spoon. Commonly an apple is eaten without an ustensil. In each
case, the actions of the lips, teeth, and tongue are different. Further variations in sense are
introduced when the agent is considered. Compare (…)
The executive ate the steak
The baby ate the steak
The dog ate the steak
Each of these sentences gives rise to different suppositions about location, circumstance,
manner, instrumentality, and antecedent and consequent conditions. The general point is that
a word could have different meaning in a very large number of the sentences in which it might
appear, even when there is core meaning”(p.168, 169)
Subtleties about improvisation can be grasped only through the analysis of the details that the
authors provide. For this reason we examine how authors describe, analyze and reflect on
improvisation. Table 8 summarizes the main aspects of diversity that are taken into account in
this work.

  111
 Table 8. Aspects of improvisation in our analysis

Stances on diversity Aspects of improvisation

Traditional aspects of diversity
Realism stance
Concept instrumentation
Theoretical foundations
Problems
Methods
Epistemological approach
Context of improvisation
Improvisers
Reasoning about improvisation
Analysis of improvisation
Description of improvisation

Section 2. Method

Our analysis followed a three steps process. First, we developed and iteratively defined a
classification grid. Second, we completed factorial and cluster analysis of the output of the
classification of articles that deal with improvisation. Third, we interpreted the statistical
outputs. The diverse steps of our analysis process are represented in Figure 21.

  112
 Figure 21. Classification and analysis process

2.I. Classification process

Classification is a widespread practice in academic life. Even unconsciously, individuals tend
to gather similar objects and separate distinct ones so as to make sense of their actions
(Bowker, Star, 2000). According to these authors, classification is « a spatial, temporal or
spatio-temporal segmentation of the world ». Therefore, a classification system is « a set of
boxes (metaphorical or literal) into which things can be put into then do some kind of work –
bureaucratic or knowledge production » (p. 10). In our view classification is the most
appropriate way to attain our objectives: i) It is a systematic approach that sets uniform
criteria of analysis for all the papers, ii) It optimizes chances to detect divergences between
studies, iii) It allows peer discussion on predefined categories as well as statistic treatment as
a basis for scientific validity.

  113
To conduct this study, we reviewed 105 articles and essays that deal with improvisation in
organizations. We selected studies in which improvisation is explicitly mentioned in the
abstract. Studies come from diverse fields such as MIS, marketing, innovation management,
organization theories and entrepreneurial literature. We used the following process to collect
papers: We first performed a first request on Business in ABI/INFORM Complete database
with the following specifications: Improvisation in the citation document text (excluding book
reviews, dissertations, newspapers) with Scholarly Journal as Publication Type from 1990 to
2010 resulted in 1882 papers9. However, this initial sample was too big so we refined our
requests. We performed a second request10 to target papers in which improvisation appeared
in the citation and abstract. We completed the results by replicating the same request on
EBSCO and Scholar Google. The screenshots of the requests and further detail on this part of
the classification process can be found in the Annexes.
We excluded some papers that mentioned improvisation in the abstract without detailing its
conceptualization, such as Wagner, Nevell and Piccoli’s work (2010) as well as Knox et al.
(2007). Finally, we used the snowball sampling technique to enrich the sample with some
studies that explicitly mentioned improvisation in the findings and that were frequently cited
in papers. We viewed them as fundamental papers that develop a complete view of
improvisation but were out of the scope of requests because the term improvisation was
missing from citation and abstract. Such a refinement concerned less than 10% of the sample.
The process of selection of such papers was iterative and led to discussion between the
authors. For instance, Lanzara (1983; 1999), Barley (1986) and Barry and Rerup (2006)
represent some of the works that were considered to be included in the sample. However,
these papers do not detail their conceptualization of this notion, and we therefore excluded
them from the sample. Annex 1 presents the articles that were selected in our classification.
We primarily collected articles that appear in widely known journals. Our rationale is that
these articles are influential within the research community and play a significant role in the
construction of theories of improvisation. We have excluded conference proceedings from our
sample because a significant proportion of conference papers eventually get published in
journals.
Two of the authors completed a cross coding to refine the definition of the variables. The
cross coding provided the opportunity to discuss divergences related to the subject of
classification between authors. Through the cross coding process, we identified three main
sources of divergence: i) Some variables were interpreted and used differently. Thus, the
authors agreed on a unique definition of these variables. ii) The authors used different
categories but agreed on their analysis of the papers, which led us to simplify the scales, iii)
the identification of some categories was subjective. We relied on Miles and Huberman’s
work (1994) to evaluate the agreement rate between authors during the cross coding. We
attained an overall agreement rate of 57%.

9
The last request was completed on the 10th of July 2010
10
The latest request was completed on the 25th of July 2010
  114
 2.II.Classification grid
The classification process complies with the following rules. First of all, classification should
rely on consistent classificatory principles, such as date. Secondly, the categories are mutually
exclusive. Thirdly, the classification should provide a fairly complete representation of the
world. In reality, the third condition remains most of the time unsatisfied.
Table 9 presents the variables and categories that compose our classification grid in this work.
We detail each of the 11 variables and the 33 categories that compose our analysis grid. We
iteratively developed the variables and the categories through the classification process.

  115
 Table 9. Classification variables and categories

Variables Description Categories

Period At what period was the study published? Period 1(1992-1998)
Period 2 (1999-2003)
Period 3 (2004-2010)
Reasoning What is the mode of reasoning about Deductive
about improvisation? Inductive
improvisation Abductive
Epistemological What are the ontological assumptions that Functionalist
approach of authors do on improvisation? Interpretive
improvisation Radical
Primary level of What organizational actors do the authors Individual
analysis observe? Group
Organizational
Contribution to What kind of knowledge do authors bring Accumulation centered
knowledge on about improvisation? Accumulation non centered
improvisation Renewing centered
Renewing non centered
Research design What research design do authors adopt? Empirical quantitative
Empirical qualitative
Conceptual
Focus on What is improvisation useful for? Design
improvisation Use
practices
Scope of What degree of instrumentality do First level
enquiry authors depict? Second level
Conditions for Is improvisation triggered or emergent ? Exogenous
improvisation Endogenous
emergence Mixed emergence
Improvisation What is the degree of uncertainty and Crisis
context time constraints when improvisation Turbulence
takes place Normal
Hierarchical What is the hierarchical status of the Low
status of improvisers? Middle
improvisers Top
Mixed
2.II.1. Period
Period refers to the period of publication from 1991 to 2010. We divided the overall period of
publication into three sub periods. Cuts between periods correspond to the release of special
issues on improvisation in academic journals. We assumed that each period distinguished
from the others because of the influence of these special issues. The first period (1992-1998)
gathers articles that were published until the release of the 1998 special issue on
improvisation in Organization Science. The second period goes from 1999 to 2003, when

  116
International Journal of Studies Management released a special issue on improvisation. The
final period consequently starts in 2004 and ends in 2010.
2.II.2. Epistemology/paradigmatic posture
We heavily drew from Burrell and Morgan’s work (1979) to assess the papers’ approach to
improvisation. We agree with Deetz (1996) on the added value of Burrell and Morgan’s four-
paradigm matrix. However, we were reluctant with classifying the authors’ epistemological
postures because these are sometimes implicit. Applying Burrell and Morgan, we examined
whether the two following assertions applied to the description of improvisation in paper: i)
the reality of improvisation is embedded in individuals’ subjectivity ii) Improvisation is
related to discontinuities, shifts and changes. We then deduced whether the authors’
epistemological posture was interpretive, functionalist or radical.
We provide some examples of our rationale with respect to this category. Zack (2000) relates
how improvisation genres developed from deviations. He depicts improvisation as a
subjective experience: “While improvisation is grounded in forms and memory (…), each
improviser must determine to what extent they want to improvise, within those forms, with
those forms, or outside those forms?” (p. 230). The reality of improvisation that he reports is
embodied in individuals’ experience and subjective. As he explains: “Ordinary conversation
is (…) more interplay than dialog, lodging people together in an intersubjective world in
which participants mutually and iteratively create meaning out of interaction” (p.232).
Conversely, the experience of improvisation gets out of the scope of enquiry of Davis et al.,
(2009). These authors adopt a functionalist view in that they focus on the influence of
structure as a distinct reality from individuals’ subjectivity. Moreover, they target predictive
truth, which suggests the existence of some environmental order. Their proposition 2 is a fair
illustration of their approach (p. 8): “As environmental dynamics increases, the optimal
amount of structure decreases”.
2.II.3. Reasoning about improvisation
We focused on the logical reasoning with respect to improvisation in studies. Authors’
discourse and rationale about improvisation may be deductive, inductive and abductive when
the authors go back and forth between empirical reality and theory in the same argument.
Mendonça and his colleagues, in their work on wild cards (2004b), investigate the properties
of improvisational capabilities in a purely deductive manner. In particular they extensively
draw on Weick (1993) and Bigley and Robert’s work on incident command system (2001) to
infer on improvisation and planning as complementary in crisis response. As they conclude:
“Putting these arguments together, it is possible to suggest that organizations need plans but
must be aware that these plans should be accommodated to local circumstances” (p.211).
Conversely, some authors rather adopt an inductive perspective on improvisation. They start
from empirical reality to convert it into abstraction concepts. This is the case of Weick’s work
(Weick, 1993, 1998). Other authors develop their thoughts abductively. For instance, Hatch
(1999) relies on both literature and personal experience of jazz to come up with properties of
organizational structures – ambiguous, emotional and temporal - and recommendations to

  117
manage them. To us, her approach is abductive: she imports properties of organizational
improvisation from the jazz metaphor that she confronts with organizational theory. On the
point of the ambiguity of structure that is an inherent feature of improvisation, she first draws
from jazz the following statement “This openness (or lack of closure) in structure permits any
of the musicians to take the tune in a variety of directions, which, if played well, contribute
innovation”. (p. 85). Then she relies on theory to enrich her reflection on ambiguity of
structure: “In organization theory, ambiguity was first conceptualized in relation to
organizational decision making and choice (…). However, whereas March and Olsen
theorized ambiguity as part of the explanation for the limits of rationality in organizational
choice processes, the jazz metaphor encourages us to reinterpret these empty spaces as
opportunities to improvise”(p. 86).
2.II.4. Levels of analysis
We drew inspiration from Markus and Robey’s work (1988) to define levels of analysis of
improvisation. Specifying the level of analysis is a frequent practice among academics when
it comes with empirical work (Mason, 2002). It has been presented as a condition of good
quality of theoretical work as well (Markus, Robey, 1988). Markus and Robey define the
level of analysis as “the entities about which theory poses concepts and relationships” and
provide four occurrences for that category: individuals, groups, organization and society
(1988). However, our sample did not include studies that analyze improvisation at the societal
level, which restricted the level of analysis to three categories. For example, Leybourne and
Sadler-Smith (2006) focus on the analysis of individuals’ intuition in projects. Sharkansky
and Zalmanovitch enhance improvisation as a peacekeeper practice in the Israel Government
(2000). As they explain (p.323): “These improvisations in effect consisted of bending the
implementation of formal policies which, in fact, could not be implemented without employing
massive force (…) and raising the conflict to an unacceptable pitch. It enabled the city to be
governed”.
2.II.5. Contribution to knowledge on improvisation
We addressed two dimensions of knowledge creation related to improvisation: i) whether
knowledge is accumulated or renewed, ii) whether authors create knowledge on improvisation
or peripheral concepts. First, theoretical contribution can consist of an assimilation or an
accommodation of knowledge (Piaget, 1977). Accumulation consists of elaborating
knowledge on the basis of previously developed definition and frameworks. On the contrary,
renewing view of notions contributes to exclude irrelevant theories. The second dimension
concerns the proximity between the concept of improvisation and the knowledge created. We
distinguished centered contribution in papers that aims to direct infer on improvisation and
non-centered contribution that concerns themes that surround improvisation. For instance,
Moorman and Miner follow a knowledge-building approach by proposing a definition of
improvisation on the basis of prior work (1998a): “We argue that collective improvisation
may be produced by the joint activities of individuals, who are themselves improvising. The
joint action of the individuals produces a “system” that we label a collective improvisation”
(p. 704). We have labeled such contribution Accumulation and centered because it provides
  118
knowledge on the very notion of improvisation in a cumulative manner. Conversely, Vera and
Crossan original insights on improvisation (2005) as their work “builds on the principles and
insights from improvisational theater to unpack the nature of collective improvisation ». They
question the widespread tenet of improvisation as a factor of increasing performance in teams.
We labeled this kind of paper renewing and centered. Some other papers that we label
renewing non centered are critical but improvisation is out of the scope of their contribution.
For instance Faia-Correia (2003) questions the legitimacy of rule obedience within teams.
Finally some authors provide more knowledge building that involves the concept of
improvisation such as Konsynski and Tiwana (2004) who elaborate on the notion of
improvisational network. We label these studies accumulation non centered.
2.II.6. Research design
We based the classification of research design on Yin’s work (2003) that enables a clear
distinction between the core designs that have been reported in literature so far. We simplified
Yin’s categories of research designs and distinguished three categories: i) Empirical
quantitative that includes any quantitative treatment in relation to some empirical reality. This
category includes surveys and simulation, such as Davis, Eisenhardt and Bringham (2009). ii)
Empirical qualitative designs refer to both longitudinal and qualitative case studies such as
Ford (2008) and Baker (Baker, 2007). Finally, iii) Conceptual works refer to studies that deal
with improvisation that rely on metaphorical and concrete examples, such as Hatch (1999).
2.II.7. Focus on improvisation practices
We distinguished authors’ interest in designing or using systems during improvisation. There
has been an extensive debate on the relevance of the conceptual distinction of design from
implementation in organizational life (Lin, Conford, 2000) and in improvisation (Baker,
2007). No one can deny that use implies some design, and reciprocally, all the more as
improvisation involves simultaneously design and use. We argue that studies focus either on
use or design, even though these practices are not independent from each other. The first
modality of this category, e.g. design, refers to articles that rather depict improvisers with
respect to the design of systems of action, such as strategies, processes, solutions. For
example, Davis and colleagues connect the notion of improvisation to strategy formulation
and focus on improvisation as a design process (2009). Similarly, Akgun and Lynn explore
new product development as improvisation practices (2002). Other studies focus on
improvisation within the use of systems. In these kinds of studies, the focus is on
improvisation that occurs during implementing activities. In other words improvisers’ do have
some plans or predefined process but the fit between this material and the situation is
ambiguous. King and Ranft, for instance, explore surgeons’ improvisation on existing
procedures to cope with emergency (2001).
2.II.8. Scope of enquiry
This category deals with the scope of investigation of improvisation by the authors. The
modality that we call “first level” refers to a specific focus on improvisation as a cognitive
effort to cope with a specific situation or problem. That modality gathers studies that

  119
primarily depict improviser’s efforts to overcome obstacles to action, due to lack of definition
or resources. The modality that we call “second level” refers to a larger view of improvisation
in studies that include in their analysis of improvisation organizational artifacts, such as
technology, objects and knowledge. The second level considers the reinvention process by
improvisers of resources, as defined by Rice and Rogers (1990, ref à intégrer). King and
Ranft’s work (2001) illustrates well a “second level” study. For these authors, improvisation
is not only a way to cope with ambiguity but also triggers the development of new
knowledge. From their observations of surgeons’ practices, they deduce the importance of
improvisation for individual or collective learning “Each surgical procedure is an application
of scientific or technical knowledge but also (…) a discovery of new knowledge that can be
applied in future procedures” (p.266).
2.II.9. Conditions of improvisation emergence
We considered two possible occurrences about improvisation emergence that are endogenous
and exogenous. We drew this distinction from current research on endogenous versus
exogenous strategy factors (Dechamp et al., 2007). The exogenous category refers to papers
that focus on improvisation as a response to some unexpected event such as a lack of resource
or a natural hazard. The endogenous category refers to studies that stress that improvisation
stems from the individuals and the groups’ tendency to deviate from established practices. For
example, Faia-Correia (2003) demonstrates the emergence of improvisation through the
evolution of practices. Similarly, Orlikowski (1996) stresses improvisation is incremental
rather than resulting from any specific event. Most likely, improvisation is not exclusively
endogenous or exogenous but covers the two dimensions. In fact, most of the time,
improvisation is a response to an event that demonstrates the inadequacy of existing plans,
processes or resources. To this extent, the conditions of emergence of improvisation are both
exogenous and endogenous. In this case, we used the mixed emergence category.
2.II.10. Improvisation contexts
Improvisation context refers to the uncertainty and time constraints that improvisers face. It
has three categories: crisis, turbulence and normality. We rely on the description of
environmental uncertainty and time constraints to classify articles. First of all, we labeled
“normal” the articles that did not mention extraordinary time constraints and uncertainty such
as King and Ranft (2001). Secondly, we used the category “turbulent” to refer to the articles
that mentioned environmental velocity such as Eisenhardt and Tabrizzi (1995). Finally, the
category “crisis” corresponds to articles that depict improvisation in the context of crisis.
2.II.11. Hierarchical status of improvisers
We used four categories to depict the hierarchical status of improvisers: low status, middle
status, high status and mixed status. We focus on the hierarchical status of improvisers that
are likely to be concerned by the articles’ findings. Mixed status refers to studies in which
authors do not refer to a specific group of improvisers. For instance King and Ranft examined
improvisation among surgeons but deduced guidelines in relation to various hierarchical
layers (2001).

  120
 Section 3. Statistical treatment

3.I. Descriptive statistics

In this section we compare the frequency of categories within the sample. Figure 25 presents
the overall frequencies of categories within the sample. First of all, the number of studies on
improvisation significantly increased from 1992-1998 to 2004-2010, which confirms that
improvisation has been attracting the attention of a growing number of researchers.
Secondly, more than 70% of the studies focus on individuals’ improvisation. In addition,
more than 60% of studies provide a second level description of improvisation. In these studies
improvisation appears as a component of organizational reinvention and change. More than
50% of studies tackle improvisation in turbulent and critical contexts, which suggests that
authors relate improvisation to uncertainty and time constraints, as suggested in previous
literature reviews (Cunha et al., 1999c).
We identify that some categories become more frequent over time. Interestingly, both crisis
and turbulence are increasingly coming to the attention of researchers who study
improvisation. However, the academics’ interest for turbulence was at its climax between
1999 and 2003. In the latest years, crisis has more attracted academic attention, as represented
in Figure 22. A majority of studies have a functionalist perspective on improvisation and the
functionalist epistemological posture has become predominant over time, as represented in
Figure 23. Similarly, deductive reasoning has become commonplace among studies, as
represented in Figure 24. In addition, conceptual studies are dominant in the sample of
articles, representing more than a third of studies. However, the frequency of conceptual
studies has remained stable from one period to another. In addition, the frequency of
quantitative studies has increased from around 15% to 25%. In our view, this statistic
confirms previous assertion that conceptual effort on improvisation has been intense (Cunha
et al., 1999c).

  121
 Figure 22. Improvisation context categories frequency over time

Figure 23. Epistemological categories frequency over time

  122
 Figure 24. Reasoning categories frequencies over time

  123
 Figure 25. Categories overall frequency

  124
 Figure 26. Categories frequency over time radar

Finally, the radar represents the evolution of categories frequencies from one period to
another. In Figure 26, Group 1 refers to articles that were published between 1992 and 1998,
Group 2 refers to articles that were published between 1999 and 2003. Finally, group 3 refers
to articles published since 2004. The radar output suggests that approaches of improvisation
are globally similar from one period to another. We deduce that despite significant evolutions
related to reasoning on improvisation and epistemological posture, the frequency of many
categories is similar from one period to another.
3.II.Principal Component Analysis
We completed the remainder of our statistical analysis in two steps. First of all, we evidenced
some factors that account for diversity within research on improvisation in the management
field. Secondly, we completed a cluster analysis to search for groups of categories. We used
factor analysis to evidence similarities among variables within the sample.
We used two techniques to make sense of the diversity of articles within the sample. We
completed a Categorical Principal Component Analysis (CATPCA) because our categories
are nominal data. CATPCA is a similar technique as PCA and is used to deal with categorical
variables whose values require quantification. In parallel, we completed a PCA on the
transformed variables, which enabled us to use the rotation algorithm VARIMAX on the

  125
 sample. Similarly to Principal Component Analysis (PCA), CATPCA can be used to detect
groups of variables that evolve similarly within the sample and to reduce the variables. The
combination of PCA and CATPCA has been regularly used in statisctics (Manisera et al.,
2006).
In the first steps of the statistical treatment, the running of the CATPCA unveiled that both
Authors’Affiliation and Improvisation ontology did not load on any dimension. We deduced
from these low values that these variables did not evolve with other variables. Some of the
other variables’ loadings did not differ from one dimension to another, which led us to
simplify the scales. For instance, we simplified the Research Design variable into three
categories (for more detail see, the presentation of variables section). By doing so, the
component loadings converged and diverged more clearly.
The PCA on the transformed data helped to detect the number of relevant components within
the sample. We transformed the data by using the principal (VPRINCIPALS) method for
quantification. We then compared the results obtained by the PCA on quantified categories.
The contribution of the PCA on quantified categories enabled to detect the number of
components that were relevant for our analysis. Both the screen test and the examination of
the total variance explained suggested that four components (or dimensions) accounted for
60,6 % of the total variance. Similarly, the CATPCA on four dimensions revealed 61% of the
variance accounted for.
Table 10 presents the eigenvalues and the variance explained from the PCA on quantified
variables. Table 11 presents the eigenvalues and the variance accounted for from the
CATPCA on four dimensions. These tables show similar results from both treatments.

Table 10. Total Variance Explained

Extraction Sums of Squared Rotation Sums of Squared

Initial Eigenvalues Loadings Loadings
Componen % of Cumulative % of Cumulative % of Cumulati
t Total Variance % Total Variance % Total Variance ve %
1 2,297 20,883 20,883 2,297 20,883 20,883 1,789 16,267 16,267
2 1,812 16,475 37,358 1,812 16,475 37,358 1,749 15,905 32,172
3 1,496 13,597 50,956 1,496 13,597 50,956 1,606 14,597 46,769
4 1,063 9,666 60,621 1,063 9,666 60,621 1,524 13,852 60,621
5 ,938 8,528 69,149
6 ,762 6,932 76,081
7 ,679 6,171 82,252
8 ,613 5,576 87,829
9 ,575 5,229 93,058
10 ,411 3,737 96,796
11 ,352 3,204 100,000
Extraction Method: Principal Component Analysis.

  126
 Table 11. Variance explained by the Model - CATPCA on Four dimensions

Variance Accounted For

Total
Dimension Cronbach's Alpha (Eigenvalue) % of Variance
1 ,629 2,335 21,226
2 ,469 1,742 15,833
3 ,353 1,472 13,383
4 ,192 1,212 11,014
Total ,937a 6,760 61,457
a. Total Cronbach's Alpha is based on the total Eigenvalue.
As the second step of our statistical treatment, we checked the reliability of our PCA analysis.
We checked that our PCA of the sample was reliable from several aspects. First of all the
Eigenvalues of the dimensions were all superior to 1. Moreover the reliability tests of
component analysis reveal that KMO and Bartlett’s test are satisfactory. The KMO index is
superior to 0.5, which shows that there is a minimum level of collinearity between variables,
which makes our PCA relevant (Jolibert, Jourdan, 2006).

Table 12. Reliability measurement of the PCA on transformed values

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. ,574
Bartlett's Test of Sphericity Approx. Chi-Square 179,662
df 55
Sig. ,000
Finally, we analyzed component loadings and compared them to the component matrix and
the rotated component matrix to identify the loadings of the variables on each of the four
components. Table 13 presents the components loadings and Table 14 presents the rotated
component matrix.

  127
 Table 13. Component loadings – CATPCA

Dimension
1 2 3 4
Period -,442 -,381 -,156 ,358
Epistemo ,545 ,477 -,204 -,133
Reasoning -,647 ,004 ,435 ,311
LevelAna -,473 ,050 -,456 -,156
ImpEmergence -,094 -,559 -,617 -,110
FocusPractice ,577 -,293 ,209 ,469
ScopeEnqu ,527 ,214 -,221 ,466
ContribKnowledge -,064 ,519 -,335 ,598
ResearchDesign ,593 -,161 -,313 -,220
Context -,151 ,702 ,202 -,283
HierStatImprov ,449 -,367 ,534 -,097
Variable Principal Normalization.

Table 14. Rotated Component Matrix - PCA on the transformed variables

Component
1 2 3 4
Periode Quantification -,094 -,399 ,638 -,089
Epistemo Quantification -,030 ,396 -,338 ,548
Reasoning Quantification -,041 -,843 -,009 -,092
LevelAna Quantification -,629 ,016 ,170 -,062
ImpEmergence Quantification -,283 ,260 ,744 ,008
FocusPractice Quantification ,794 ,134 ,170 ,095
ScopeEnqu Quantification ,392 ,252 ,038 ,538
ContribKnowledge -,008 -,073 -,112 ,800
Quantification
ResearchDesign ,136 ,718 ,084 -,013
Quantification
Context Quantification -,402 -,162 -,627 ,175
HierStatImprov Quantification ,581 ,165 -,242 -,482
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 7 iterations.
3.III. Cluster analysis
To detect similarities between studies, we examined the proximity between the categories. To
do so, we used hierarchical cluster analysis techniques, as detailed by Hair and his colleagues
(Hair et al., 1984). More specifically, we have extracted the factorial scores of each modality,
e.g. the coordinates of the modalities in the referential composed by the four dimensions.

  128
Our cluster analysis on the set of 33 categories did not enable us to detect any significant
group. We used the replication technique to validate our findings. The replication technique is
recommended by Aldenderfer and Blashfield (1984, p. 65) and consists of replicating the
same cluster analysis with various algorithms to verify that the dendogram patterns are stable
across the various techniques used to form clusters (Durrieu, Valette-Florence, 2005). Annex
4 provides some example of the similarities between our replicated dendogram. In Figure 27,
we represent the dendogram that we obtained with the Ward’s method. In this dendogram, the
category “crisis” is an outlier, e.g. a category with low proximity with other categories. In
addition, the dendogram reveals that 5-10 iterations aggregated only a small number of
categories. We conclude from these findings that our sample does not include any
homogenous group of category

  129
 Figure 27. Categories overall dendogram

We then used cluster analysis techniques on the fragments of categories that loaded on each
dimension. We present these clusters in the following sections.

  130
 Section 4. Interpretation of findings

The PCA and CATPCA techniques unveiled four statistical components. We interpret these
components as four dimensions of diversity in the management literature on improvisation.
We label the four dimensions as follows: i) Evidencing, ii) Focalizing iii) Contextualizing and
iv) Creating knowledge. We also view the dimensions as meta-variables in our classification
as we deduced these dimensions from the variables loadings in the CATPCA and PCA
outputs. Going further in our interpretation, we propose that each dimension corresponds to a
specific task in the research process.
We then considered each component. We performed a cluster analysis on the portions of
categories that loaded on each component. For each component, we identified two groups
from the cluster analysis. We interpret this finding in the following way. With respect to each
task, e.g. evidencing, focalizing, contextualizing and creating knowledge, two primary
academic perspectives coexist. Making the sum of two perspectives per each task, we identify
eight perspectives on improvisation that are not exclusive. Figure 28 presents the four
dimensions and the eight perspectives on improvisation.

Figure 28. Eight perspectives on improvisation

First of all, authors observe different kinds of improvisers and improvisation practices. They
focus on different groups of improvisers and distinct improvisation practices. While some
authors focus on improvisation among top managers in designing activities, other studies

  131
explore how individuals from lower or mixed hierarchical layers improvise when they
implement or use resources.
Secondly, authors conceptually and empirically evidence improvisation in distinct manners.
Some authors deduce characteristics of improvisation from preexisting frameworks while
others develop their own conceptualization of improvisation.
Thirdly, the context of improvisation differs across studies. A specific stream of research
focuses on crisis improvisation to respond to unanticipated events. Other papers of the sample
study improvisation as a practice that progressively emerges and structures organizational
practices.
Finally, research on improvisation differs on the basis of the knowledge created. While some
papers target a clarification of improvisation in a functionalist perspective, another stream of
research evidence relates improvisation and organizational functioning in a more critical or
interpretive fashion. Table 15 presents the various dimensions of diversity and the related
variables.
Dimensions of diversity within the Variables
management literature on
improvisation
Dimension 1: Evidencing Reasoning about improvisation
Research design
Dimension 2: Focalizing Level of analysis
Focus on improvisation practice
Hierarchical status of improvisers
Dimension 3: Contextualizing Context for improvisation
Improvisation emergence
Period
Dimension 4: Creating knowledge Epistemological posture
Scope of enquiry
Contribution to knowledge

Table 15. Dimensions of diversity and corresponding variables

4.I. Evidencing improvisation

Evidencing corresponds to the methodological techniques that authors employ to recognize an
organizational phenomenon as an improvisation.
This dimension refers to three possible paths of reasoning that authors use to define
improvisation as well as make theoretical assumptions and propositions. Some authors rely on
a theoretical source to provide a definition of improvisation and it’s characteristics. For
example, Lau and his colleagues rely on Orlikowski and Hoffman’s improvisational model to
identify and understand improvisation practices into clinical settings (Lau et al., 1999). These
authors use an existing model as a conceptual foundation to recognize and discuss
improvisation. Some other studies primarily rely on empirical observation to conceptualize

  132
improvisation. By doing so, they either induce their own view of improvisation or discuss
potential existing theories in an abductive manner. For example, Hatch draw on both her
experience of jazz music to question traditional understandings of organizations and propose
alternative ways to understand ambiguity and emotionality in organization (Hatch, 1999). By
doing so, she proposes an alternative view of improvisation and organizations.
Evidencing also refers to the techniques used to identify and analyze improvisation, e.g. the
research design. Most quantitative studies develop or rely on preexisting measurement tools
to analyze improvisation. For instance, Moorman and Miner provide a three items scale to
measure organizational improvisation level of actions (Moorman, Miner, 1998b, p. 10).
Conceptual studies provide a rich description of past improvisation and detail the authors’
rationale for differentiating and diverse types of improvisation. For example, Weick explains
that improvisation can be more or less intense and describes different levels of intensity of
improvisation (1998). To illustrate this point, we present Lanzara’s work (1999) that specifies
improvisation as a component of bricolage:
« We observe a variety of mundane building activities, recombinations of preexisting
components, small scale practical experiments, local readjustments and repairs,
extemporaneous improvisations, which taken together resemble what, by using a French
word, can be labeled bricolage (Lévi-Strauss, 1962) » (p. 334).
As suggested by the dendogram, two main perspectives can be identified as divergent
techniques to provide evidence of improvisation. First of all, quantitative studies, such as
surveys, introduce improvisation in a deductive fashion. For instance, Eisenhardt and Tabrizzi
(1995) relies on previous work to evidence the role of improvisation in organizations that
handle uncertainty. These authors then deduce the role played by improvisation in the
experiential model of firms’ fast adaptation to turbulent environments. Secondly, qualitative
research induces improvisation characteristics from in-depth analysis or metaphors. Even
though they inspire from and remain consistent with organizational theories, the authors from
this stream of thought develop original conceptualizations of improvisation. For instance,
Clegg, Cunha and Cunha draw inspiration from a dialectical approach to organizations to
recognize and define improvisation as a paradox between acting and planning (2002).
Similarly, Ciborra proposes an original conceptualization of improvisation as an inherent
practice in organizing (1999):
“While improvising, the agent is able to frame and recombine features of her situation, so that
they become resources for intervention. In a burst of action the contours of the problematic
situation, plans for problem solving and the deployment of resources coalesce. Improvisation
is intentional but extemporaneous, that is, happening almost unexpectedly (‘ex tempore’—
outside the flow of time), and with little known cause or relationship”(p. 78).

  133
 Figure 29. Evidencing dendogram

4.II.Focalizing
The second dimension that we label focalizing corresponds to the authors’ focus when they
observe improvisation in organizations. Focalizing refers to three variables. First of all, the
hierarchical status of improvisers responds to the following question: who is improvising
among organizational members? Secondly, the variable entitled practices of improvisation
correspond to the question of what is the object of improvisation? Finally, the level of
analysis describes who is observed within the organization.
We identify two main manners of observing improvisation in organizations, as suggested by
the dendogram. First of all, some authors specifically deal with improvisation within top
management. For instance, David, Eisenhardt and Bingham deal with strategic improvising
that consists of the improvised design of strategy (2009). They focus on top managers and
decision makers as actors of improvisation at the organizational level.
These articles depict improvisation as collective output. Conversely, some other studies
explore improvisation as a practice among all sorts of actors within the organizations. While
some authors provide information about the status of improvisers within the organizations,
some others are less specific. Instead of targeting on a specific hierarchical status, they focus
the individuals’ experience of improvisation when they use organizational means and
resources.

  134
 Figure 30. Observing dendogram

4.III. Contextualizing
The third dimension that we label contextualizing refers to the empirical context of
improvisation. Contextualizing refers to three variables. First of all, it refers to the level of
uncertainty and time pressure that improvisers face. Some studies examine improvisation
when uncertainty and time constraints are not inherent to the organization’s environment. For
instance, Orlikowski examines how technology users progressively develop local
improvisations to handle the misfit of technology to their needs (1996). She hardly refers to
uncertainty as a condition of emergence of improvisation but primarily focuses on the
sociotechnical environment of users as one of components of improvisation. Similarly,
Macredie and Sandom find a correlation between local improvisation and customers’
dissatisfaction in normal and turbulent environments (1999). Some other studies pay more
attention to organizational contingency and take into account turbulence as an ingredient to
improvisation. For instance, Brown and Eisenhardt highlight improvisation as a response to
environmental uncertainty (1997). Finally, some authors focus on crisis improvisation.
We rely on the contextualizing dendogram to distinguish two main manners of
contextualizing improvisation. A first approach consists of viewing improvisation as way of
organizing that enables organization to deal with more or less turbulence. A second stream of

  135
research investigates improvisation as an exogenous phenomenon. The dendogram reveals
that studies that examine how organizational members react to crisis through improvisation
have been more frequent since 2003. For instance, Webb examines role improvising as a
technique to cope with uncertainty regarding resources and tasks (2004). Roux-Dufort and
Vidaillet find that the occurrence of improvisation depends on individual’s perception of
crisis (2003a). Contextualizing refers to the period of completion of the study as well. As our
findings suggest, the academic interest for organizational context can depend on the period of
time when the study is completed. Because some topics are more fashionable at certain times
among academic audience, crisis, as a context of improvisation, has been increasingly
investigated.

Figure 31. Contextualizing dendogram

4.IV. Creating knowledge

The fourth and final dimension that we label creating knowledge relates to the type of
knowledge that authors create on organizations. Including the authors’ epistemological
posture, this dimension unveils whether authors propose that knowledge on improvisation has
some predictive power (functionalist perspective). Conversely, some authors assume that
knowledge on improvisation is not univocal given that improvisation has changing attributes

  136
depending on individual’s subjectivity (interpretive perspective) or environmental and
ideological shifts (radical perspective). The creating knowledge dimension also refers to the
scope of investigation of improvisation. The first level studies narrow investigation to the
emergence and occurrence of improvisation. However, second level studies provide a wider
view of the organizational implications of improvisation by relating improvisation to other
concepts such as innovation and organizational change. Finally, creating knowledge refers to
the type of knowledge created in the sample studies.
We distinguish two main perspectives on knowledge creation on improvisation. First of all,
some authors build on previous work to accumulate or unify knowledge on improvisation as
an objective and predictable phenomenon. For instance, Cunha, Cunha and Kamoche (1999c)
provide an integrative definition of improvisation from existing definitions. By doing so, they
combine some distinct aspects of improvisation such as bricolage and extemporaneity into
their conceptual propositions. Contrasting with this perspective, other studies aim to offer
original perspectives on improvisation as an organizational phenomenon. For instance, Faia-
Correia (2003) investigates how improvisation can reverse the established hierarchical order
within an organization. By doing so, she offers new and controversial perspectives on the
influence of improvisation on organizational functioning.

  137
 Figure 32. Knowledge creation dendogram

Section 5. Discussion of findings

In this paper, we detail our interest in handling diversity in research on improvisation. We

argue the need not to restrict our analysis to theoretical criteria such as theoretical foundation
or research questions. On the contrary, we suggest that diversity in research on improvisation
stems from at least two interrelated sources: the empirical reality and how authors depict,
analyze and reflect on this reality.
Following this approach, we have classified the literature on improvisation by considering
various aspects of improvisation. Our findings from the PCA, CATPCA and cluster analysis
reveal that diversity does exist because we could not identify any predominant approach of
improvisation. Furthermore, our findings update Cunha, Cunha and Kamoche’s work on
research on improvisation (Cunha et al., 1999c). Conceptual papers remain predominant, but
empirical methodologies are increasingly used. Our work reveals temporal evolutions of
research on improvisation regarding epistemological postures, reasoning and improvisation
contexts. In addition, we suggest that diversity of research on improvisation stems from
divergences between authors in relation to four primary research tasks: i) evidencing, ii)
focalizing, iii) contextualizing and iv) creating knowledge on improvisation foster diversity of
research on improvisation.
The contribution of this work to academic life is twofold. First, we confirm that research on
improvisation is effectively diverse. To explain this diversity we submit three hypotheses. A
first hypothesis would be that this diversity is due to the specificities of the concept of
improvisation. Considering other topics, one can realize that research on topics such as ICT
fit (Zigurs, Khazanchi, 2008), technology diffusion (Fichman, 1992), knowledge (Alavi,
Leidner, 2001) or even culture (Gallivan, Srite, 2005), converge around a limited set of
perspectives. However, further investigation of improvisation is necessary to confirm or
infirm this hypothesis. An alternate explanation would be that research on improvisation is
too “young” to structure into a unified field of research. Improvisation has started attracting
researchers’ attention for two decades on the contrary to other concepts that have been studied
for a longer time. The final explanation is related to our holistic approach to diversity. By
taking into account a large panel of variables, we enhanced heterogeneity within our sample.
We also identify the major differences between existing approaches of improvisation in the
management literature. Thus, this study offers some criteria to distinguish studies that
approach improvisation in a similar manner from others. This techniques has already been
used to provide clarity in research domains (Prescott, Conger, 1995). Authors can compare
studies’ common points and differences with respect to recognizing, observing,
contextualizing and creating knowledge. As a result, two studies that seemingly have nothing
in common can be compared to promote knowledge building. For example, our study reveals
that Akgün and Lynn’s (2002) approach of improvisation is similar to Chelariu, Johnston and
Young’s work (2002). At first glance, these two studies deal with unrelated topics: while the

  138
former study investigates improvisation in New Product Development (NPD) teams, the latter
focuses on organizational learning from improvisation. However, by using our research
outputs as an analysis grid, one can find that these two studies offer complementary findings
and views of improvisation. Both these studies adopt the same approach to improvisation as a
collective phenomenon in top hierarchical layers. Furthermore, both studies depict
improvisation as an inherent practice in organizations and provide consistent guidelines on
how to develop improvisational capabilities. Akgün and Lynn recommend that managers
promote their team’s stability over time (2002). Chelariu and colleagues go further by
highlighting the importance of stable interactions and wisely shared information flows within
the team (2002). Similarly, we recommend that authors specify their approach of
improvisation at each of the four steps that we describe in this study to promote knowledge
building and discussion among researchers.
In spite of all our efforts, some limits and criticism can be raised regarding our work. First of
all, one can easily question the relevance of classification in research. Given its exclusive
theoretical basis, classification implies lack of experience and can lead to simplistic
conclusions (Sandelands, Srivatsan, 1993). Deetz (1996) argues that the temptation to reduce
rich paradigms or conceptions to simplistic categories is commonplace in research. Moreover,
there is the risk of bias because of our overusing the “other” category. By doing so, some
insightful and original papers on improvisation can be marginalized and eventually
overlooked (Deetz, 1996). In addition, As Bowker and Starr suggest (2000), classification is
rarely complete, in spite of completeness being an essential feature of classification in theory.
Finally, classifying implies the risk to extrapolate or misinterpret the authors’ view on
improvisation. Consequently, we had important discussion on categories in order to determine
whether they could be retained for statistical treatment.
Secondly, our scope of inquiry is not exhaustive. This work aims to address diversity in
research on improvisation. However we have eluded in this work conceptual polysemy. One
major aspects of polysemy of improvisation is that some other surroundings concepts are used
instead of improvisation, such as bricolage or tinkering. It would be interesting to create a
conceptual mapping to represent the logical connections between improvisation and other
concepts. Similarly, we only depict mainstream or at least well-known papers because we
assume that these will have a significant impact on current and future research. It is worth
knowing more about conference proceedings and their influence in knowledge building.
Given our findings and the limitations of our work, we propose three avenues for future
research. Our findings partly suggest that research on improvisation is diverse and includes
more than 8 different perspectives. Why does a unifying perspective on improvisation not
exist? One option to respond to this question would be to further trace the development of the
research field on improvisation. Thus the next step to this work is to account for the evolution
of research on improvisation year after year. Using k-means cluster analysis, one could
develop a historical view of the development of research on improvisation, therefore
identifying some influential stream of research more specifically.

  139
Another direction for research could be to enrich existing work on diversity by addressing
concepts’ polysemy in management. Polysemy refers to the plurality of meanings possibly
taken by a concept (Deane, 1988; Pisanelli et al., 2004) and promotes a diverse understanding
of concepts within the same field. Undeniably, polysemy also results from diversity in that it
stems from the use of various terms to depict similar phenomena. Concepts such as strategy
and values are relevant illustration of rampant polysemy in management (Hambrick, 1980).
For instance, the term value has been widely used in radically different settings. For some
authors in strategic management, value may simply mean performance or at least the creation
of additional benefit thanks to an optimal organizational function. Porter’s thought (1991) is
one of the most thorough developments on the notion of value, notably through the concept of
value chain. In a single glance, one can intuitively acknowledge that the notion of value refers
to a wide panel of managerial realities, coming from profitability to satisfaction.
Our final direction for future research consists of exploring the connections between
improvisation and peripheral concepts such as bricolage, resilience, ambidextry, tinkering etc.
Linkages between improvisation and other concepts are diverse but comprise mostly causal-
effect, inclusion or mean-end relationships. For example, improvisation has been depicted as
a means for organization to be more resilient by Rerup (2001). In fact resilience was defined
as an organization’s ability to deal with its vulnerabilities (McManus, 2008) and results not
only from anticipatory but also improvisational skills (Rerup, 2001). Our literature review
also suggests empirical investigation is required on such relationships in that authors have
developed contradictory views on improvisation role in organizational life. For instance,
Bansler and Havn (2004) suggest improvisation enables sensemaking. Lanzara (1999), on the
contrary, posits improvisation prevents sensemaking.

Conclusion

To conclude this study, we propose to use our findings to position our doctoral work. Our
evidencing of improvisation follows the reflexive approach. In chapter 2.3 we discuss existing
conceptualizations and their limitations to identify the need to identify the role played by
interactions in organizational improvisation during crisis response. We then make theoretical
propositions that we deductively draw from existing literature. We then complete an in-depth
qualitative analysis of improvisation. With respect to focalizing, we diverge from the two
main perspectives on improvisation that we document in this study by focusing on
improvisation as a collective phenomenon between mixed hierarchical layers. We
contextualize improvisation as a response to critical events. With respect to knowledge
creation, we primarily attempt to offer additional insights on interactions that compose
organizational improvisation and ICT use. Our objective is to propose an alternative view of
organizational improvisation. Figure 33 represents our position in relation to the four
dimensions and the eight perspectives that we document in this study.

  140
 Measurement and
Reflexive
deduction

Evidencing

Improvisation as a
Individuals' collective process Top management
improvisation to that involves mixed improvisation as a
cope with resources hierarchical status collective design

Observing

Improvisation as a way Improvisation as a

of organizing response to crisis

Contextualizing

Offering alternative
Knowledge building
views

Creating knowledge

Figure 33. Our position in relation to improvisation in the dissertation

  141
 Annex 1. Description of the sample

Table 16. Description of the articles sample

Year &
SamplingTechnique Authors and reference Journal Name
Bastien, Hostager, Cooperation as
communicative accomplishment: A
1992 symbolic interaction analysis of an Communication
Snowball Sampling improvised jazz concert Studies
1993 Weick, the collapse of, Man Gulch
Snowball Sampling Disaster ASQ
Spark, Machiavellianism and personal Journal of the
success in marketing: The moderating role Academy of Marketing
1994 REQUEST of latitude for improvisation Science
Eisenhardt, Tabrizi, accelerating adaptive
1995 process: product innovation in the Global
Snowball Sampling Computer Industry ASQ
Orlikowski, Improvising organizational
transformation over time: a situated
1996 REQUEST change perspective ISR
Crossan, Lane, White, The improvising
organization: Where planning meets Organizational
1996 REQUEST opportunity Dynamics
Brown, Eisenhardt, The art of continuous
1997, REQUEST change ASQ
International Studies of
Aram, Walochik, Improvisation and the management and
1997 REQUEST Spanish manager organization
Berniker, Working the jazz metaphor:
1998 REQUEST Musings driving down I-5 past midnight Organization Science
Barett, Creativity and improvisation in
jazz and organizations: Implications for
1998 REQUEST organizational learning Organization Science
1998 REQUEST Crossan, improvisation in action Organization Science
Moorman, Miner, The convergence of
planning and execution: Improvisation in
1998 REQUEST new product development Journal of Marketing
Moorman, Miner, organizational
1998 REQUEST improvisation and organizational memory AMR
Weick, Improvisation as a Mindset for
1998 REQUEST analysis. Organization Science
Accounting,
management, and
Ciborra, Notes on improvisation and time information
1999 Snowball Sampling in organizations technologies
1999 REQUEST Cunha, Cunha, Kamoche, Organizational International Journal of

  142
 improvisation: What, when, how and why Management Reviews
Lanzara, Between transient constructs and
persistent structures: designing systems in Journal of Strategic
1999 Snowball Sampling action Information Systems
Lau et al., Patterns of improvisation for Information
1999 REQUEST evidence based practice in clinical settings technology and people
Hatch, Exploring the Empty Spaces of
Organizing: How improvisational Jazz
1999 REQUEST halps redescribe organizational structure Organization Studies
Macredie, Sandom, It-enabled change:
1999 REQUEST Evaluating an improvisational perspective EJIS
Sharkansky, Zalmanovitch, Improvisation
in Public administration and policy Public Admnistration
2000 REQUEST making in Israel. Review
Dyba, Improvisation in small software
2000 REQUEST organizations IEEE Software
Zach, Jazz improvisation and organizing -
2000 REQUEST Once more from the top Organization Science
Bigley, Roberts, The Incident Command
System: High-reliability organizing for
2001 REQUEST complex and volatile environments AMJ
Cunha, Cunha, The brave paradoxical
2001 REQUEST world Strategic Change
Kamoche, Cunha, Minimal structures:
From jazz improvisation to product
2001 REQUEST innovation Organization Studies
King, Ranft, Capturing knowledge and
knowing through improvisation: what
managers can learn from the thoracic Journal of
2001 REQUEST surgery board certification process Management
Miner, Bassoff, Moorman, Organizational
2001 REQUEST improvisation and learning: A field study ASQ
Rerup, "Houston, we have a problem":
Anticipation and Improvisation as sources Comportamento
2001 Snowball Sampling of organizational resilience organizacional e gestao
Sashittal, Jassawalla, Marketing
implementation in smaller organizations: Journal of the
Definition, framework, and propositional Academy of Marketing
2001 inventory Science
Akgyn et al., Multi-dimensionality of European Journal of
learning in new product development Innovation
2002 REQUEST teams Management
European Journal of
Akgun, Lynn, New Product Development Innovation
2002 REQUEST Team Improvisation and speed to market Management
Chelariu, Learning to Improvise, Journal of Business
2002 REQUEST Improvising to Learn Research
2002 Snowball Sampling Clegg, Cunha, Cunha, Dialectic View Human Relations
2002 REQUEST Heeks, Information Systems and The Information

  143
 Developing Countries: failure, success, Society
and Local Improvisation
Lewis et al., Product development
tensions: Exploring contrasting styles of
2002 Snowball Sampling project management. ASQ
McGinn, Keros, Improvisation and the
logic of exchange in socially embedded
2002 REQUEST transactions ASQ
McKnight, Bontis, E-improvisation:
collaborative groupware technology
expands the reach and effectiveness of Knowledge and
2002 REQUEST organizational improvisation Process Management
Journal of Business
2002 REQUEST Pava, The path of moral growth. Ethics
Tsoukas, Chia, On Organizational
Becoming: Rethinking Organizational
2002 Snowball Sampling Change Organization Science
Baker, Miner, Eesley: Improvising Firms,
2003 REQUEST Bricolage, Research Policy
Cunha, Kamoche, Cunha, Organizational
Improvisation and Leadership. A field Internation Studies of
Study in Two Computer Mediated managemet and
2003 REQUEST Settings organization
Cunha, Cunha, Organizational Journal of
Improvisation and Change: two Syntheses Organizational Change
2003 REQUEST and a filled gap Management
Creativity and
Cunha, Gomes, Order and Disorder in innovation
2003 REQUEST Product Innovation management
International Studies of
Faia-Corriea, Mind the gap between management and
2003 REQUEST processes and practices organization
Internation Studies of
management and
2003 REQUEST Kamoche, Improvisation in organizations organization
Kamoche, Cunha, Cunha, Towards a
Theory of OI Locking beyond the The Journal of
2003 REQUEST metaphor of jazz Management Studies
Montuori, The complexity of
improvisation and the improvisation of
2003 REQUEST complexity Human Relations
Pinnington, Morris, Pinnington, The
Relational Structure of Improvisation-A International Studies of
Case Illustration from Corporate Video management and
2003 REQUEST Production organization
Roux-Dufort, VIdaillet, The Difficulties International Studies of
of Improvising in a Crisis Situation-A management and
2003 REQUEST Case Study organization
Cunha, Organizational time: a dialectical
2004 REQUEST view Organization
  144
 Ibott, O'Keefe, transforming the Ericsson-
2004 REQUEST Vodafone Relationship Long Range Planning
Konsynski, Tiwana, The improvisation-
efficiency paradox in interfirm electronic
networks governance and architecture Journal of Information
2004 REQUEST considerations Technology
Mendonça, Cunha, Kaivo-oja, Ruff,Wild
Wards, Weak Signals and Organizational
2004 REQUEST Improvisation. Futures
Mendonça, Wallace, Studying International Journal of
organizationally-situated improvisation in Mass Emergencies and
2004 REQUEST response to extreme events. Disasters
International Journal of
Webb, role improvising during crisis Emergency
2004, Snowball Sampling situations Management
Stolegraaf, Dickson, The paradox of a Joural of the Academy
2004 REQUEST marketing planning capability of Marketing Science
Vera, Crossan, Theatrical improvisation:
2004 REQUEST Lessons for organizations Organization Studies
Chédotel, L'improvisation
organisationnelle, Revue Française de
2005 REQUEST Gestion RFG
Crossan, Cunha, Vera, Cunha, Time and Academy of
2005 REQUEST Organizational Improvisation Management Review
Gauthereau, Hollnagel, Planning, Control, European Management
2005 REQUEST and Adaptation Journal
Pinnington, Learning in a competitive Internation Journal of
field: MBA students' improvised case Human Resource
2005 REQUEST studies of IHRM Management
Purser, Petranker, Unfreezing the future:
Exploring the dynamic of time in The Journal of Applied
2005 REQUEST organizational change Behavorial Science
Vera, Crossan, Improvisation and
2005 REQUEST innovative performance in teams Organization Science
Akgun, Lynn, Yilmaz, Learning process
in new product development teams and Industrial Marketing
2006 REQUEST effects on product success Management
Brigham, Introna, hospitality,
improvisation and gestell: a Journal of Information
2006 REQUEST phenomenology of mobile information. Technology
Cunha, Cunha, Towards a complexity
2006 REQUEST theory of strategy Management Decision
Duguid, What Telking about Machines
2006 REQUEST tells us Organization Studies
Elbanna, The validity of the improvisation
argument in the implementation of a rigid Journal of Information
2006 REQUEST technology Technology
Heeks, Health Information Systems: International Journal of
2006 REQUEST Failure, success and improvisation Medical Informatics

  145
 Hmlieleski, Corbett, Proclivity for
improvisation as a predictor of Journal of Small
2006 REQUEST entrepreneurial intentions Business Management
Johanson, Johanson, Turbulence,
Discovery and Foreign Market Entry: A
Longitudinal Study of an Entry into the Management
2006 REQUEST Russian Market International Review
John, Grove, Frisk, Improvisation in Managing Service
2006 REQUEST service performances: lessons from jazz Quality
Leybourne, Managing change by
abandoning planning and embracing Journal of General
2006 REQUEST improvisation Management
Leybourne, Improvisation within the
project management of Change, Some Journal of Change
2006 REQUEST observations from UK Financial Services Management
Leybourne, Sadler-Smith, The role of
intuition and improvisation in project International Journal of
2006 REQUEST management Project Management
Mendonça, Friedrich, Training for
improvisation in emergency management: International Journal of
opportunities and limits for information Emergency
2006 REQUEST technology Management
Akgun, Byrne, Lynn, Keskin, New Journal of Engineering
Product Development in Turbulent and Technology
2007 REQUEST Environments Management
Baker, Resources in play: Bricolage in the Journal of Business
2007 REQUEST Toy Store (y) Venturing
Dennis, Macaulay, using jazz to
investigate improvisation and market
2007 REQUEST orientation EJM
Faia Correia, Cunha, Getting Started
Initializing Organization around new
2007 REQUEST technology Management Research
Jambekar, Improvisation model for team
performance enhancement in a Team Performance
2007 REQUEST manufacturing environment. Management
Journal of
Mantere, Sillince, Music as a metaphor for Organizational Change
2007 organizational change. Management
McDaniel, Management strategies for Performance
complex adaptive systems: Sensemaking, Improvement
2007 REQUEST learning and improvisation Quarterly
Mendonça, Decision support for
improvisation in response to extreme
events: Learning from the response to the Decision Support
2007 REQUEST 2001 World Trade Center attack System
Ieee Transactions on
Mendonça, Wallace, A cognitive model of Systems Man and
2007 REQUEST improvisation in emergency management Cybernetics
2007 REQUEST Mendonça et al., Collaborative Communications of the
  146
 adhocracies and mix-and-match ACM
technologies in emergency management
Vera, Rodriguez, leading improvisation: Organizational
2007 REQUEST lessons from the american revolution Dynamics
Carlsson, ElSawy, Managing the five Information systems
tensions of IT-enabled decision support in and e-business
2008 SnowballSampling turbulent & high velocity environment management
Bergh, Lim, Learning how to restructure:
Absorptive capacity and improvisational
views of restructuring actions and
2008 REQUEST performance SMJ
Erden, Krogh, Nonaka, The quality of Journal of Strategic
2008 REQUEST group tacit knowledge Information Systems
Ford, Complex adaptive systems and
improvisation theory: Toward framing a Journal of Change
2008 REQUEST model to enable continuous change Management
Hmlieski, Corbett, the contrasting
interaction effects of improvisational Journal of Business
2008 REQUEST behavior Venturing
Samra et al., Effects of improvisation on
product cycle time and product success: A
study of new product development (npd) International Journal of
2008 REQUEST teams in the united states Managemet
Bingham C, Oscillating improvisation: Strategic
How entrepreneurial firms create success Entrepreneurship
2009 REQUEST in foreign market entries over time Journal
Daly, Grove, The impact of improvisation
training on service employees in a
2009 REQUEST European airline: a case study EJM
Davis, Eisenhardt, Bingham, Optimal
Structure, Market Dynamism, and the
2009 REQUEST Strategy of Simple Rules ASQ
Journal of Information
Kautz, Improvisation in Information Technology Case and
2009 REQUEST Systems Development Practice Application
Leybourne, culture and organizational Journal of service
2009 REQUEST improvisation in UK financial services science management
Int. Journal of
Leybourne, improvisation and APM: a Managing Projects in
2009 REQUEST comparative Business
Oakes, Freedom and constraint in the
2009. REQUEST empowerment as jazz metaphor. Marketing Theory
Vendelø, Improvisation and Learning in
Organizations: An opportunity for future Organizational
2009 REQUEST empirical research Learning
Yanow, Tsoukas, What is reflection-in- Journal of
2009 REQUEST action? A phenomenological account Management Studies
Di Domenico, Social Bricolage:
Theorizing Social Value Creation in Entrepreneurship:
2010, REQUEST Social Enterprises Theory & Practice
  147
 Mendonça, Webb, Burst, L'improvisation
dans les interventions d'urgence: les
relations entre cognition, comportement et
2010 REQUEST interactions sociales Tracés

  148
Annex 2. Description of the requests

Figure 34. Request screenshot

An example of the second type of request on GoogleScholar – Improvisation in Title, resulted

in 195 papers (similar request was led on ABI/Inform – with the citation AND abstract
option- resulted in 116 papers and EBSCO – Improvisation in Abstract OR Title- resulted in
126 papers).

  149
Annex 3. Statistical outputs

Table 17. Number of iterations during the CATPCA

Iteration History
Variance Accounted For Loss
Restriction of
Centroid to
Iteration Centroid Vector
Number Total Increase Total Coordinates Coordinates
0a 6,232624 ,000005 37,767376 37,153416 ,613961
1 6,343096 ,110472 37,656904 37,153416 ,503488
2 6,470576 ,127480 37,529424 37,062312 ,467111
3 6,549961 ,079385 37,450039 37,014874 ,435164
4 6,604051 ,054090 37,395949 36,984251 ,411698
5 6,642576 ,038525 37,357424 36,961684 ,395740
6 6,670698 ,028122 37,329302 36,943885 ,385417
7 6,691536 ,020837 37,308464 36,929473 ,378991
8 6,707147 ,015612 37,292853 36,917723 ,375130
9 6,718960 ,011813 37,281040 36,908135 ,372905
10 6,727984 ,009024 37,272016 36,900314 ,371702
11 6,734939 ,006955 37,265061 36,893934 ,371127
12 6,740343 ,005403 37,259657 36,888729 ,370928
13 6,744567 ,004224 37,255433 36,884485 ,370948
14 6,747884 ,003318 37,252116 36,881028 ,371087
15 6,750499 ,002614 37,249501 36,878218 ,371284
16 6,752562 ,002063 37,247438 36,875938 ,371500
17 6,754192 ,001630 37,245808 36,874096 ,371712
18 6,755480 ,001288 37,244520 36,872611 ,371910
19 6,756497 ,001017 37,243503 36,871419 ,372085
20 6,757299 ,000803 37,242701 36,870465 ,372236
21 6,757932 ,000633 37,242068 36,869705 ,372363
22 6,758431 ,000499 37,241569 36,869101 ,372468
23 6,758824 ,000393 37,241176 36,868624 ,372552
24 6,759133 ,000309 37,240867 36,868248 ,372619
25 6,759376 ,000243 37,240624 36,867954 ,372670
26 6,759568 ,000191 37,240432 36,867724 ,372708
27 6,759718 ,000150 37,240282 36,867547 ,372735
28 6,759836 ,000118 37,240164 36,867410 ,372754
29 6,759929 ,000093 37,240071 36,867306 ,372764
30 6,760002 ,000073 37,239998 36,867228 ,372769
31 6,760060 ,000057 37,239940 36,867171 ,372769
32 6,760105 ,000045 37,239895 36,867129 ,372766
33 6,760140 ,000036 37,239860 36,867100 ,372759
34 6,760168 ,000028 37,239832 36,867081 ,372750
35 6,760190 ,000022 37,239810 36,867070 ,372740

  150
36 6,760208 ,000017 37,239792 36,867064 ,372729
37 6,760221 ,000014 37,239779 36,867062 ,372717
38 6,760232 ,000011 37,239768 36,867064 ,372704
b
39 6,760240 ,000008 37,239760 36,867068 ,372691
a. Iteration 0 displays the statistics of the solution with all variables, except variables with
optimal scaling level Multiple Nominal, treated as numerical.
b. The iteration process stopped because the convergence test value was reached.

Table 18. Variance accounted for each variable

Table 19. Communalities

Initial Extraction
Periode Quantification 1,000 ,583
Epistemo Quantification 1,000 ,572
Reasoning Quantification 1,000 ,720
LevelAna Quantification 1,000 ,428
ImpEmergence Quantification 1,000 ,702
FocusPractice Quantification 1,000 ,687
ScopeEnqu Quantification 1,000 ,508
ContribKnowledge 1,000 ,659
Quantification
ResearchDesign 1,000 ,541
Quantification
Context Quantification 1,000 ,612
HierStatImprov Quantification 1,000 ,657
Extraction Method: Principal Component Analysis.

  151
 Table 20. Component transformed matrix

Component 1 2 3 4
1 ,609 ,681 -,242 ,327
2 -,464 -,103 -,639 ,605
3 -,499 ,495 ,617 ,353
4 ,406 -,529 ,391 ,634
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.

  152
Annex 4. An example of replication
We provide in this annex an example of the outputs that we obtained by replicating cluster
analyses with diverse algorithms. This example corresponds to our replication of our cluster
analysis of the focalizing dimension categories. From one dendogram to another the groups
are the same, despite small variations in relation to distance.

Figure 35. Between-groups linkage dendogram

  153
 Figure 36. Highest neighborhood dendogram

Figure 37. Centroid method dendogram

  154
Chapter 2.3 Information and Communication Technology ((ICT)),
crisis, and organizational improvisation: theoretical propositions and
literature review

Introduction

As previous chapters explain, crisis has increasingly become a subject of extensive research,
and a primary concern for managers. Industrial accidents are more likely to happen due to
increasing organizational and technical complexity (Perrow, 1984), and organizations remain
vulnerable to acts of terrorism and natural disasters. A significant body of work in crisis
management focuses on the prevention of crises.
Because of organizations’ inability to prevent or avoid accidents and disasters, crisis response
has become an equally important aspect of crisis management, especially managing the
immediate aftermath to limit additional consequences. During crisis response, planning and
anticipation are necessary (Rerup, 2001; Shrivastava, 1993). However resilience is crucial
because one must expect reality to be different from what was anticipated (Rerup, 2001;
Waugh, Streib, 2006).
Crisis response requires organizations to engage in unfamiliar actions without the time to
formulate detailed plans. Depending on the nature of a crisis, organizations may need to
revise previous plans, when there are no guidelines that can be clearly followed or when the
situation is unexpectedly new and complex. Every crisis is unique and planning for crisis
response is limited by the uncertainty inherent in the notion of crisis (Billings et al., 1980).
For these reasons, improvisation is considered as a core element of crisis response
(Mendonça, 2007). The organizational ability to improvise is crucial to avoiding rigidity-
threat effects during crisis response (Staw et al., 1981).
For this reason, researchers have led great efforts to analyze improvisation from both
cognitive and behavioral perspectives. For example, Cunha and his colleagues elaborate a
generic definition of improvisation, followed by an exhaustive classification of
improvisation’s treatment in academic literature (Cunha et al., 1999c). A significant amount
of work has been achieved to move the focus of analysis of improvisation to the
organizational level. For example, Moorman and Miner’s work (1998a, b) is representative of
the prominent studies that aim to delineate improvisation as a measurable construct. These
authors have contributed to a better knowledge of the antecedents of organizational
improvisation such as turbulence, real time information and organizational memory
(Moorman, Miner, 1998b). However three specific aspects of improvisation as an
organizational response to crisis remain unclear.
First of all, the mechanisms by which routine practices support improvisation remain
undefined. The degree to which improvisation departs from routine is variable, which leaves
unanswered the general question of the connection between improvisation and routine.
Additionally, in the context of crisis, routines can transform into a threat for organizations

  155
(Barnett, Pratt, 2000; Staw et al., 1981), which makes it necessary to clarify the role of
routine in crisis improvisation.
Secondly, the question of how improvisers manage to coordinate with each other requires
further understanding. Widely accepted definitions of improvisation emphasize the short
delay between decision and action but fall short of details on interactions that permit
extemporaneous coordination. For instance, Moorman and Miner predict that heavier
information flows contribute to the emergence of improvisation (Moorman, Miner, 1998a).
However, their results lack details on how improvisers avoid risks of information overload
and select relevant information while using ICT tools.
Thirdly, ICTs’ reliability for supporting improvisation remains uncertain. The role of
information and communication technology (ICT) in crisis management has also been
investigated (Comfort, 1993; Quarantelli, 1997). ICTs provide crucial support for
communication and coordination during disasters. In particular, mobile devices enable fast
communication and the prevention of additional damage. When examining crisis
improvisation, one can infer that real time communication fits well to the need for a limited
delay between decision and action that features improvisation (Moorman, Miner, 1998b).
However, ICTs have been criticized for wasted time caused by information overload (Hiltz,
Turoff, 1985; Quarantelli, 1997, 2007), the lack of user-friendly tools (Mendonça, 2007) as
well as integration issues (Dawes et al., 2004).
This study seeks to fulfill two related needs: 1) proposing an explanation of how collective
improvisation works and 2) investigating the role that ICTs play in improvisation during crisis
response. The first need exists because some of the literature on organizational improvisation
provides rich but disconnected findings, thus making it necessary to provide an overview of
the mechanisms that structure organizational improvisation. We begin by analyzing previous
definitions of organizational improvisation in section 1. In section 2, we make some
theoretical propositions regarding the constituents of organizational improvisation. We then
address the second need by reviewing the literature on ICTs support to crisis improvisation.
Section 3 presents our method and section 4 presents our findings. We finally discuss the
limitations of our work in section 5.

Section 1. Theoretical background on improvisation

This section details our analysis of the literature on organizational improvisation. We first
describe some aspects of improvisation unveiled by existing definitions. By discussing the
limitations of these definitions, we emphasize the need for further understanding of some
specific aspects of organizational improvisation in crisis response.
1.I. Improvisation definitions
Table 21 summarizes eight different ways in which organizational improvisation has been
conceptualized. Although all the definitions refer to innovative responses, there is little in
common across definitions. However, the first four definitions capture the relationship
between improvisation and planned activity. By describing improvisation as a situation in
  156
which the time between the conception of an action and its execution is reduced, Moorman
and Miner (1998a) characterize improvisation as “decision as action unfolds”. This means
that organizational members make decisions as they are already in the process of enacting
them. The need to improvise is also associated with unstable and high-velocity environments
in which innovation is necessary (Crossan et al., 2005 Cunha, 2005; Eisenhardt, 1989).
Organizations may be compelled to improvise when there is “time pressure or emergency to
respond to an unexpected problem” (Weick, 1993). Still, organizations do not always
improvise in these circumstances, such as during crises (Roux-Dufort, Vidaillet, 2003a).

Table 21. Definitions of improvisation

Source Definition
(Ciborra, 1996a) “Thinking and action emerge simultaneously and on the spur of
the moment”
(Vendelø, 2009) “Convergence between acting and planning, and reworking of
pre-composed material in relation to unanticipated ideas”
(Moorman, “Organizational improvisation refers to the convergence of
Miner, 1998a) conception and execution”
(Berliner, 1994, “Composing and performing extemporaneously, conception as
cited by (Bansler, action unfolds”
Havn, 2004)
(Lanzara, 1999) “Unexpected practices, deviation from routines, puzzling or
random behaviors calling for interpretation”
(Meyer, 1998) “Devising resourceful solutions to intractable problems”.
(Preston, 1987) “A process of introducing changes to an initial model while
maintaining continuity in the performance”
(Crossan, 1998) “Action is taken in spontaneous and intuitive fashion”
(Cunha et al., “The conception of action as it unfolds by an organization and/or
1999c) its members drawing on available material, cognitive, affective
and social resources”
The second five definitions in Table 1 all refer to resources. The idea that improvisation
involves the creative recombination of resources at hand, also known as bricolage (Cunha et
al., 1999a), is fundamental to these definitions. Particularly in crisis situations, organizations
do not have the time or capacity to acquire new resources. Improvisation is unlike
organizational invention in that it departs from what the organization accomplished and
learned from the past. For Rerup (2001), improvisation is a form of resilience, which is
achieved by recombining past experience and resources into new patterns of action. Cunha,
Cunha and Kamoche (1999c) refer to the wide range of material, cognitive, affective and
social resources that may be recombined to generate improvised responses to crisis.
1.II.Some conceptual limitations
Some theories oversimplify the connection between structures and improvisation as an
antinomy. Undoubtedly less structure provides more room for agility and improvisation
(Davis et al., 2009). Furthermore, improvisation serves as an alternative way of functioning

  157
when planning impossible or irrelevant. For example, in a high velocity environment,
improvisational strategic decision making enables organization to react to threats and
opportunities (Perry, 1991a). Similarly, Weick argues that improvisation is merely a negation
of foresight and planning (1998). However, many examples of crisis indicate that operational
routines are the only means to complete response when planned strategy cannot be followed.
By opposing improvisation to routine, scholars imply that effective management involves the
abandonment of plans and structures because they inhibit creative responses. In our view, the
conceptual ties between improvisation and routine require further investigation.
In fact, the simultaneous occurrence of planning and acting in improvisation forms an
organizational paradox (Clegg et al., 2002; Joffre et al., 2006). It leads us to notice that the
paradoxical nature of acting and planning mirrors the simultaneity of ends and means which
is logically and temporally contradictory. How can individuals coordinate their actions as they
develop coordination mechanisms? How can plans and actions occur simultaneously? These
paradoxes illustrate the tensions inherent in improvisation while they also beg for more
analytical thought about how such paradoxes can be resolved (Lewis, 2000).
How to shed light on these uncertainties regarding improvisation? One could draw from the
jazz metaphor to understand the ins and outs of improvisation interactions. Most definitions
of organizational improvisation draw upon the analogy between management action and
artistic performance, usually in jazz or theatre (Cunha et al., 1999c; Vera, Crossan, 2005).
From the jazz metaphor, extemporaneous improvisation is understood as intentional
variations on melody, harmony or rhythm to produce music that is fresh and new. These
metaphors reflect the inherent complexity of improvisation, which can be compared to the
complexity of managerial responses to crisis situations. They suggest that managers can act
similarly to creative artists by improvising instead of following prescribed plans. Although
the insights drawn from jazz and theatre metaphors are useful (Hatch, 1999; Kamoche,
Cunha, 2001; Vendelø, 2009), they are limited by the differences between art and business
contexts (Hatch, 1998). Even though management is a purposeful activity, its objective is not
creative expression. Moreover, the crises managers are faced with in the business world differ
from the artistic risks taken by the jazz or theatre performer. Additionally, Organizations are
more complex than even the largest jazz orchestra or Wagnerian opera, so improvisation in
organizational contexts cannot simply mirror the principles derived from the study of
performing arts. ICT is also not typically used to assist improvisers on stage.
Even though opportunities to observe organizational improvisation live are scarce, we argue
the necessity to pursue investigation on how improvisation functions. We believe that
improvisation has been all too often presented as a magical solution to the inability to plan
during a crisis. An exclusively positive view of improvisation ignores potentially harmful
consequences in which safety and security may be threatened by abandoning standard
procedures. Improvisation can pose some challenges for organizations such as decreased
effectiveness (Davis et al., 2009; Moorman, Miner, 1998b), loss of organizational cohesion
(Miner et al., 2001) and interpersonal conflicts (Aram, Walochik, 1996). Many treatments of
improvisation overstate the importance of spontaneity at the expense of misunderstanding the

  158
relationship between spontaneous and planned action (Vera, Crossan, 2005). Identifying what
is improvisational in organizations is a starting point to propose mechanisms by which
managers can successfully manage the occurring improvisation.
From our review of the literature on improvisation, we identify two needs. First of all, since
organizational improvisation is a collective response, the coordination between improvisers
needs greater emphasis. Furthermore, a more satisfactory analysis of the interactions that
compose organizational improvisation should reconcile the paradox of planning and action as
simultaneous activities. In other words, organizational improvisation theory should articulate
the relationship between plans and actions more precisely. Secondly, the manner in which
available resources are combined should be specifically addressed. ICTs should be included
as a resource so that their potential contribution to crisis response is not ignored. These
clarifications would provide a more practical understanding of organizational improvisation,
one from which actionable responses may be derived. In the following section, we articulate a
more comprehensive description of organizational improvisation, drawing from a variety of
sources of knowledge on improvisation.

Section 2. Organizational improvisation constituents

In our work, we view organizational improvisation as a collective process that involves

spontaneous deviation from established uses of resources and that requires coordination
among actors who command resources. We propose to analyze improvisation not as an
isolated event but rather as a component of a wider process of organizational adaptation. This
analysis implies that improvisation is a crucial factor for long-term stability and builds on
organizational capabilities and resources.
In this section, we propose a description of the functioning of organizational improvisation.
From our analysis of the literature on crisis management and improvisation, we have deduced
five main constituents of organizational improvisation. We define each of these constituents
and then explain their role in organizational improvisation. This description relies on
deductions from theories. Although it is not exhaustive it offers a structured exploration of
organizational improvisation in crisis response.
Table 22 summarizes the constituents of organizational improvisation in crisis response and
their functions.

  159
 Table 22. Crisis organizational improvisation constituents and their functions

Organizational Function
improvisation
constituents
Spontaneity Understanding whether predefined procedures are executable.
Development of new ideas.
Adjustment to coworkers’ actions
Deviation from Adaptation of the ongoing course of action to the specificities of the
established use or situation
resources Development of new capabilities
Enrichment of existing capabilities
Boundary spanning Information translation
Development of a common referential: narration, information,
practices, vocabulary
Common narration of past actions
Common orientations for future actions
Heedful interactions
Integration of outside intelligence
Expertise Selection of ideas
leadership Dialogical reasoning
Development of new ideas
Integration of outsider intelligence
Minimal structures Rules for interactions and actions
Anticipation of what can be done or not
Change detection
2.I. Collective process
Organizational improvisation is collective due to the fact that it involves a coordinated
response rather than an individual response. Although individual actors contribute improvised
responses during crises such efforts, no matter how heroic or creative, can easily dissipate
without the efforts of others. Considering improvisation to be collective does not imply that
an organization acts as a unified entity; rather, organizational action involves the coordinated
efforts of many individuals representing different interests within the organization. A unified
response by an organization is rare under normal conditions and highly unlikely during crisis.
We also conceive of improvisation as a process, which occurs over a short period of time.
Thus, the study of organizational improvisation should document the sequences of events
during the aftermath of a crisis. Methodologically, this can be achieved more practically by
reconstructing events from interviews and documented sources. In the longer term, successful
improvisations may be retained as part of an organization’s repertoire, from which it may
draw from in future crises. Hatch (1998) and Weick (1998), for example, regard improvised
actions as part of the organizational memory after the crisis has ended. However, improvised
actions in the short term may not necessarily be retained for future situations requiring
improvisation. Indeed, the new repertoire may not be useful unless there is a very similar kind

  160
of crisis in the future. Given the nature of crisis, we are skeptical of the claim that improvised
responses to one crisis would always be useful for a subsequent crisis.
2.II.Spontaneity
Among the features of organizational improvisation, there is the spontaneous reaction to
unexpected events.
Many definitions emphasize the role personality traits play in spontaneous action (Webster,
1992). However, a group’s ability to formulate “an immediate development on a topic”
(George, 1997; Webster, 1992) plays a significant role in crisis improvisation. For example,
the ability to find new ideas depends on how reflection on work practices is managed (Zollo,
Winter, 2002). George and Jones (1997) identify forms of spontaneity in organizations among
which helping coworkers, protecting the organization, and making constructive suggestions
are crucial activities in organizations that improvise during crises (Larson et al., 2006).
In spite of being inherent to resilience, spontaneity is not the unique characteristic of
improvisation (Crossan, 1998). However, spontaneity develops in at least three manners in
improvising groups. First of all, improvising individuals have to spontaneously recognize
whether predefined procedures are executable. Secondly, improvisers have to develop a
mixing and matching of planned procedures (Mendonça et al., 2007). In this study we
propose that improvisers have to quickly adjust to coworkers’ actions to participate in the
ongoing improvisation. We therefore propose that spontaneity during improvisation does not
only concern idea generation but coordination as well.
2.III. Deviation from established uses of resources
The second feature of organizational improvisation is the creative treatment of resources by
actors. Improvisers deviate their use of resources from established patterns during crisis
response to meet the necessity to urgently complete some tasks whose devoted resources are
not at hand (Crossan et al., 2005; Cunha et al., 1999c). Drawing on available resources rather
than seeking the proper ones corresponds to the notion of bricolage, defined by Levi-Strauss
(1962). While some authors argue that improvisation does not necessarily lead to the use of
bricolage (Baker, 2007), most studies assert that deviation from established use of resources
is an important part of improvisation (Moorman, Miner, 1998a).
Then one can wonder what role existing patterns of action play in improvisation. The largely
unresolved paradox discussed earlier addresses the requirement that improvisation include
simultaneous performance of plans and actions. Rather than regarding plans as irrelevant to
improvisation, we regard established use of organizational resources, ranging from procedures
to technology resources (Jefferson, 2006b), to be part of organizational improvisation. We
support this claim with recent research on routines and earlier work on scripts, both of which
involve resources.
Pentland and Rueters (1994) define routine as a “set of possible patterns enabled and
constrained by a variety of organizational, social, physical and cognitive structures from
which organizational members enact particular performance”. For some authors, routines

  161
serve to constrain organizational members to believe that activities that made the organization
successful in the past remain the legitimate course for the future. However, this narrow view
of routines obstructs fresh thinking and inhibits creative deviations from past practices
(Crossan, Sorrenti, 2002; Milburn et al., 1983b; Moorman, Miner, 1998a). Routine is
inherently improvisational for Feldman and Pentland (2003), who view routines not as
restricted cognitive frames but rather as generative systems that enable adaptation and
organizational change. In other words, familiar routines are renewed whenever they are
performed (Feldman, 2000). This view challenges the position that distinguishes between
improvisation and routine (Moorman, Miner, 1998a).
Routines are formed around the use of organizational resources. The rationale for describing
routines as a part of improvisation is that they are not a single repetitive pattern of actions
(Cohen, 1994; Pentland, Rueter, 1994), but rather a set of interdependent sequences of
actions. Routines are not restricted to groups or subunits, but confront different units of work,
cultures and communities of practice. Routines support intra-organizational cohesion, even in
crisis, providing a reference point from which adjustments or radical changes can be made.
By comparing improvisations with established routines, actors can gauge the degree to which
their deviations depart from established uses of resources. In this sense, improvisation
literally fills the gap between repetition of a routine and a completely original performance
(Suchman, 2007). Without knowledge of routines, actors would be unprepared to execute
spontaneous actions that differed from established ways to use resources.
Routines bear similarity to the concept of scripts. Scripts may be either individual or
collective in nature. Gioia and Poole (1984) consider scripts to be “stored knowledge that is
called into play whenever the situational cues evoke an expectation for certain events to
occur.” Although scripts are defined differently from routines, they represent a structured set
of cognitive elements that help the understanding of new situations. Scripts provide not only
an unconscious reference but also means for adaptation and improvisation. Crises are
typically the kind of situation in which scripts can assist improvisation. According to Gioia
and Poole (1984), “Other situations (than stereotypical) entail some variation on the
protoscript and require some means of distinguishing knowledge of these variations in
memory.”
By tying the definition of improvisation to the deviation from established use of resources, we
invite further comparison with the discourse on dynamic capabilities. Dynamic capabilities
correspond to an organization’s ability to satisfy environmental requirements by recombining
resources to develop capabilities that previously existed only potentially. As a response the
limitations of the resource-based-view of the firm (Barney, 1991; Grant, 1996), dynamic
capabilities theory focuses on the outcome of organizational and cognitive processes that
integrate novel applications of existing capabilities (Teece et al., 1997). Although no clear
consensus exists on the definition of dynamic capabilities, the emphasis of most definitions is
upon the ability of organizations to reconfigure and redeploy resources is different ways
(Eisenhardt, 2000; Zollo, Winter, 2002). This capability to use existing resources in novel
ways corresponds to the notion of bricolage from the improvisation literature. Zollo and

  162
Winter’s (2002) definition of dynamic capability as “a learned and stable pattern of collective
activities through which the organization systematically generates and modifies its operating
routines in pursuit of improved effectiveness” closely reflects the spirit of the improvisational
nature of resource usage.
2.IV. Interaction between actors
The third dimension of organizational improvisation in crisis response is the need for
coordination and interactions. As in most coordinated efforts, the transmission of information
is crucial to integrate a new idea or an initiative in collective functioning. Without
information transmission, subunits may have a different perception of the situation, and
improvisation may not be coordinated (Roux-Dufort, Vidaillet, 2003a). Subunits need to
remain informed about what is new and how it affects their own work, which compels them to
pay extra attention to events and listen in their coworkers (Kristensen et al., 2006; Pettersson
et al., 2004) By analogy, performers in a jazz band must listen to band mates to detect
unexpected changes introduced by another member. Normally, coordination in jazz is
accomplished with familiar signals, such as a glance, nod or a quick breath. Organizational
communication also needs to be frequent and accurate during improvisation.
Three main types of mechanisms appear to regulate interactions during organizational
improvisation: boundary spanning, expertise leadership, and minimal structures.
The first type, boundary spanning, helps information translation from one community of
users to another, as defined by Wenger (1999). Boundary spanning enables a translation of
information from one group to another to assure consistent action, thanks to the boundary
spanners that can be either individuals, such as victims (Kristensen et al., 2006), or objects
such as a technological tool (Gasson, 2005). For instance, social media and Internet web sites
enable crisis responders to publicly share information with other crisis responders and victims
(Lang, Benbunan-Fich, 2010). Different communities or subsystems may have divergent rules
or codes for action which, in addition to political tensions, can have a negative impact on the
organizational ability to react during crisis (Beamish, 2002). As crisis responses frequently
involve groups with specific rules and codes, boundary spanning is crucial to maintaining a
common referential composed of information, practices and vocabulary, or a common idea
between the individuals involved in the ongoing improvisation of how tasks were performed –
also called narration (Brown, Duguid, 1991) and should be pursued (Kapucu, 2006).
Furthermore, boundary spanning compels actors to understand that the groups they have to
collaborate with have developed different practices and values. Gasson explains that boundary
spanners often have to convince groups that their practices are not the best or need to be
adjusted, reducing the risk of misunderstandings that can endanger improvisation (2005).
Secondly, expertise helps to identify which interpretation of reality is legitimate. For example,
in high reliability organizations that have to organize instantaneously, new issues for
collective action are first addressed to experts (Weick, Sutcliffe, 2001). For this reason,
expertise leadership can regulate interactions to coordinate organizational improvisation.
Especially in crisis situations, dialogical reasoning, which consists in dialogue and feedback

  163
between experts, helps to guide the ongoing searching for ad hoc solutions (Faraj, Xiao,
2006). The emergence of expertise leadership requires that experts can contact and interact
with other participants. Expertise leadership can develop thanks to boundary objects. For
instance, social media allows discussion between interlocutors to integrate outside
intelligence and expertise into the ongoing improvisation (Turoff et al., 2004).
Thirdly, interactions that compose improvisation are based on minimal structures, such as
rules or coordination patterns, facilitate information sharing (Pinnington et al., 2003 2003).
Minimal structures include rules and codes that assure coordination and information
feedback. They are particularly important during jazz performance (Kamoche, Cunha, 2001).
Extended interpretation of the jazz metaphor reveals insights into collective improvisation and
the role that minimal structures play in the coordination among performers. For instance,
minimal rules may guide the order of solos in a performance, and signals between performers
can cue changes in the direction of the performance. However, minimal structures are also
apparent in organizational contexts.
Our reading of dynamic capabilities theory suggests that all these mechanisms participate in
the organizational adjustments that are crucial during improvisation (Zollo, Winter, 2002).
The primary purpose of systematically defining the constituents of improvisation is to guide
our review of the literature on ICTs’ role in crisis response. In the following section we
briefly describe the method for reviewing and classifying this literature.

Section 3. Method

We searched for articles that dealt with crisis response through web databases, including
EBSCOhost, Web of Science, and ABI Inform. We searched for papers that included the
keywords technology, crisis and response in the citation or abstract.11 We did not mention
improvisation in the keywords because some parts of improvisation can be described in the
authors’ work without being explicitly mentioned. We included the term response to collect
papers that provided enough practical details about actions and interactions that were
undertaken during crisis responses that were under study. The scope of selection using these
keywords went well beyond the IS literature, yielding around 110 articles including
conceptual papers, essays, and case studies. The initial screening of these papers resulted in
many not being retained either because they did not provide any evidence of improvisation,
because they did not deal with ICTs’ contribution to crisis response or because they did not
provide enough details about crisis response. To complete our search, we examined
publications from specialized journals on crisis response that were not included in the
databases, such as IJISCRAM. 12 of the remaining papers are used in this study as the basis

11
We repeated the following request SEARCH FOR full text and scholarly peer reviewed journals WITH
technology AND crisis AND response IN citation & abstract. Our last request on the 20th of August resulted in
42 papers on ABI/Inform, 31 results on EBSCO Host, 50 results on Scholar Google, with extended search in
Business, Administration, Finance, and Economics, Engineering, Computer Science, and Mathematics, Social
Sciences, Arts, and Humanities.

  164
for investigating the relationship between ICT features and constituents of improvisation.
These articles are presented in Table 25 in the Annex.
Each article’s content was extracted and entered into a summary table. We considered each
characteristic of improvisation presented in this paper and analyzed how they were supported
by ICT. For example, Table 23 shows the entry for the article “Learning from Crisis” (Dawes
et al., 2004), which reports on organizational uses of technology to cope with the loss of
information resources following the 9/11 attacks on the World Trade Center. These tables are
useful in supporting a narrative review of each article.

  165
 Table 23. Article extraction entry example

Improvisation Spontaneous Deviation from Boundary Minimal Expertise

Characteristic action established use of Spanning Structures
resources
Description Immediate Adaptive use of Using GIS, Data Emergency
creation of a applications different coordination management
data analysis initially designed groups could and applications
centre for weather follow the integration managed by
(Whittaker, forecasting to evolution of failed. There experienced
Sidner). spread information information was no volunteers
Spontaneous about availability needs and unique from other
disposal of of housing and transfer principle to states, vendor
military other resources. missing govern data experts, and
devices. Reuse of DNA information collection IT experts
software by police to the units in and quality from police.
to identify victims. need. control. The
Use of existing Resulted in management
investments with GIS data of the
other sources to problems. deployment
build an ad hoc was shared
GIS. according to
skills.
Analysis: ICT The DAC Use of standardized GIS ICT helped Many-to-
support was and modular centralized to gather many
consistent applications data for data and communicati
because all facilitated their different transmit it to ons
resources integration and groups and everyone but supported
were innovative use. enabled did not coordination
immediately Existing information provide any from
available or investment could sharing. regulation of different
recovered. be used thanks to Graphical the sources of
Internet immediate share of representatio information expertise.
network was existing resources. n provided format.
the only one unique
reliable geographical
infrastructure reference and
during the facilitated
disaster. information
sharing.
We have completed manual analysis because the number of articles we have classified has
been reduced to 12. From the entire set of articles, we identified six ICT properties that
contribute to crisis response:
(1) Graphical representation refers to an organized projection of colours and shapes on the
computer interface. Graphical representation usually inspires from cartography techniques but
can involve videos and photos as well. Conceptual and geographical mapping relies on the

  166
representing techniques. Geographic Information Systems (GIS) relies on the graphical
representation property. Other projects of common operating picture use graphical
components to provide a share representation of some crisis settings (Landgren, 2007).
(2) Modularity refers to the ability of quickly combining and reusing applications even in a
short notice. Therefore modularity supports the development of new patterns of interactions.
For example, the relatives of the victims of the hurricane Katrina combined the Google Maps
application with other programs in an ad hoc fashion to share information about the damaged
houses and victims. Even though modularity is not restricted to applications but concerns
system configuration (Turoff et al., 2004), we focus on applicative modularity in this work.
(3) Calculation supports data selection, scenario simulation and complex algorithmic
treatment of data. Artificial Intelligence (AI) extensively relies on this property (Yuan, Detlor,
2005). A certain stream of research on crisis response reports common use of the calculation
property to run simulations (Mendonça et al., 2006), or even support decision making and risk
assessing.
(4) Many-to-many communication refers to the ability of more than two users to share
information. In this way, ICTs significantly differ from phones that primarily enable bilateral
rather than multilateral exchanges. For example, many-to-many communication simply
requires clicking on “Respond to” or “Forward to” when individuals use emails to interact.
(5) Centralization of data affects the ability to access different data through one interface;
Centralization is not feasible without a preliminary organization of data and processes. For
instance, databases can represent enormous amounts of data with unique attributes so that
users can retrieve information with limited manipulation of the database during the search.
(6) Virtuality refers to the non-material characteristics of data, which facilitates recovery,
replication, modification and exchange of information. As a polysemic term, virtuality has
been the object of extensive discussion (Schultze, Orlikowski, 2001). Most researchers focus
on virtual organizations or virtual teams that correspond to group of collaborators
geographically distant from each other (Griffith et al., 2003; Kirkman, Mathieu, 2005;
Panteli, Fineman, 2005). Another stream of research has studied the features of virtual
interaction as opposed to face-to-face interaction (Massey et al., 2003; Warkentin et al.,
1997). While most studies approach virtuality as a way of interacting, we propose to define it
as a property of ICT. Virtuality enables crisis users to freely modify or update information
without any notice to collaborators. While in some cases crisis responders have the right to
modify information depending on their prerogatives or their status, in other cases every user
can potentially modify the content on online websites. This implies the risk of
misinformation. However, it enables actors to immediately update information on their own
rather than relying on someone else to transmit information.
These properties were then cross-tabulated against the five components of our definition of
improvisation. We recognized other ICT properties during analysis, such as data integration,
automation, or dynamic content. However, these were either redundant or appeared much less
frequently. So we excluded them from the cross-tabulation.

  167
 Section 4. Results

The result of cross tabulation is shown in Table 24. We draw five main propositions from our
analysis.
First of all, the graphical representation of data during crisis response creates a unique
reference that places shared information into a uniform format (Comfort, 1993). A uniformed
format for information plays two roles. First of all, converting information into the format
required by graphical codes makes it necessary for users to translate or reformulate their
ideas. In that sense technology becomes a boundary object. Secondly, graphs have the ability
to represent complex data in a way that can be comprehended with a single glance. This
feature enables fast interpretation and foster, in our view, spontaneous generation of ideas or
spontaneous adjustment.
Secondly, the combination of the centralization of data and the calculation is a starting point
for expertise leadership. For instance, intelligent systems rely on databases to select and
contact experts who are the most competent for different kinds of skills (Yuan, Detlor, 2005).
By doing so, experts have the opportunity to participate in the ongoing improvisation by
suggesting new ideas or at least providing guidelines. The design of Control and Command
systems (C2) provides another example of how ICTs can support expertise leadership. C2
systems warn experts of the need for their intervening, thanks to the centralization of data
about ongoing processes and the calculation of the gap between the course of action and the
planned response (Ntuen et al., 2006). Similarly, Decision Support Systems potentially
support dialogical reasoning between experts by detecting when the course of action deviates
from established use of resources (Mendonça, 2007).
Thirdly, combined virtuality and many-to-many communication supports both expertise
leadership and deviations from established uses of resources. First of all calculation facilitates
virtual discussions between experts enable appropriate responses (Dantas, Seville, 2006).
With no immediate cost or delay, expert can share their ideas or provide feedbacks. By
supporting discussion and dialogical reasoning, the combination of virtuality and many-to-
many communication increases the odds for a combination of different resources (Dawes et
al., 2004).
Fourthly, the structured centralization of data enables actors to find information concerning
others’ past actions, therefore promoting boundary spanning interactions and the development
of a common referential. In fact actors can build a common narrative that operates as a
translator between groups and frames rules for action (Majchrzak et al., 2007). This finding is
consistent with previous research that recommends the use of logs to trace ongoing crisis
response (Turoff et al., 2004). By just a simple click, a log can reveal a collection of
information related to past and ongoing actions.
Fifthly, coordination relies on virtuality. Boundary spanning practices between distant groups
cannot exist without immediate and free replication and exchange of information (Comfort,
1993 ). Virtuality also helps to keep shared information consistent in time, thanks to

  168
information feedback. Therefore, minimal structures can be initiated and managed by experts
(Yuan, Detlor, 2005). This helps to avoid potentially disastrous misunderstandings (Larson et
al., 2006).

  169
 Mechanisms Expertise Minimal structures Boundary Spanning Deviation from Spontaneous

 
supported by ICT leadership established use of action
resources

Graphical Comfort, 1993. Dawes et al., Palen et al., 2007.

Representation 2004. Mendonça et al., 2007. Mendonça et al., 2007.

Modularity Turoff, Chumer, Dawes et al., 2004. Dantas, Comfort, 1993.

Van de Walle, Seville, 2006. Mendonça et Dawes et al., 2004.
Yao, 2004 al., 2007. McKinney, 2008. McKinney, 2008

Calculation Yuan, Detlor, Carver, Turoff, 2007. Yuan, Detlor, 2005. Comfort, 1993.
2005. Majchrzak et al., 2007 Mendonça et al., 2007.
Comfort, 1993. Comfort et
al., 2001.
Many to many Calloway, Keen, Comfort, 1993. Yuan,
Communication 1996. Detlor, 2005.
Dawes et al., 2004. Comfort et al., 2001.

Data Dantas, Seville, Palen et al., 2007. Comfort, 1993. Mendonça et al., 2007.

Table 24. Cross tabulation results

Centralization 2006. McKinney, 2008 Palen et al., 2007. Comfort et al., 2001.
Ntuen, Balogun, Ntuen, Balogun, 2006.
2006. Turoff, Chumer, Van de Walle,
Yao, 2004

Virtuality Yuan, Detlor, Yuan, Detlor, 2005. Dantas, Seville, 2006. Comfort, 1993. Calloway, Yuan, Detlor, 2005.
2005. Mendonça Mendonça et al., 2007. Keen, 1996. Palen, Liu, Hiltz,
et al., 2007 Carver, Turoff, 2007. Dawes et al., 2004. 2007.
Majchrzak et al., 2007 Yuan, Detlor, 2005. Carver, Turoff,
Ntuen, Balogun, 2006. Dantas, Seville, 2006. 2007.
Mendonça et al., 2007. McKinney, 2008

170
Carver, Turoff, 2007.
Majchrzak, Jarvenpaa,
Hollingshead, 2007.
 Section 5. Discussion

Our objective in this paper is to provide a generic and well-argued description of

organizational improvisation and ICT support to organizational improvisation in crisis
response. Our work is exclusively exploratory and theoretical. It extends previous work by
Quarantelli (2007) and Turoff and his colleagues (2004). However this study presents some
unavoidable limitations that we detail.
First of all, our work is limited by our selective and incomplete use of prior literature. 15
articles represent a small sample. For this reason, our results require further consolidation or
at least some confrontation with empirical testing.
Secondly, we are aware of the incompleteness of the suggested description of organizational
improvisation. For example, it does not cover collaboration, which is a crucial need during
crisis (Mishra, 1996). In our view, collaboration is of primary importance during
improvisation in that it signifies that coworkers trust each other enough to improvise together.
Undoubtedly, the coordination mechanisms that we suggest in this study are of limited
interest if actors are not mentally involved in the ongoing improvisation. However, there is no
clear evidence that collaboration is a constituent of improvisation. One can argue that the
willingness to collaborate may grow between improvisers because of the ongoing
improvisation. Because of this ambiguity, we restricted our reflection to coordination. In
addition, our objective in this work was to propose a resolution of the inner paradoxes of
organizational improvisation that mainly include the question of coordination.
Similarly, our study overlooks organizational needs or constraints that are also important
during crisis response and can burden organizational improvisation. Definitely, the two
aspects of collaboration and organizational constraints should be examined in future research
on organizational improvisation.
This study focuses on technology features and does not take into account the contingencies of
the users. Crisis response implies the use of diverse tools. Considering technology at large
enables us to determine whether technological investments are useful or not. In our study,
technology is seen as an objective tool. Other perspectives suggest that technology and users
are dependent on each other. Brighan and Introna (2006) provide an interesting example of
how technology is defined by its users practices. It illustrates the limitation of our findings
that only suggest potential sources of support from technology in crisis improvisation.
“When a consultant takes a mobile phone, the consultant acquires a certain capability (to
contact and be contacted) but that is not all that happens. We need to look at this seemingly
obvious statement of ‘consultant, ‘mobile phone’ and ‘taking up’ more closely. The mobile
phone only becomes a mobile phone when taken up by the consultant. When it lies on the
table it is a potential to be ‘a mobile phone’, but it only becomes an actual ‘possibility for
contacting’ when it is picked up and one dials the number, and, of course, there is sufficient
credit on the account, and it is possible to get a signal, and so forth. In taking up the mobile
phone, both the mobile phone and the consultant get transformed. The mobile phone is no

  171
longer ‘merely’ an object and the consultant becomes a human that embodies the possibility
to contact and to be contacted at a distance”. (p. 141)
Empirical investigation may reveal further properties of technology that support
improvisation. Moreover, a more exhaustive review may strengthen our results, or it may
contradict them. Our hope is that our analysis will help guide more thorough and conceptually
sound investigations of ICTs’ role in improvised organizational responses to crises.
Our interest in how technologies can provide support to organizational improvisation remains
partial. In fact, technological support to improvisation may be achieved in other manners. For
instance, simulation systems can provide training before the emergency (Mendonça, Fiedrich,
2004). Similarly this study does not take into account mobility as an ICT property that
definitely plays a role in improvisation by enabling continuous tracking of victims (Yuan-
Hsiang et al., 2004).
Finally the scope of the inquiry of this paper is limited. Our theoretical propositions primarily
stem from our analysis of crisis management literature. As a result, these propositions focus
on crisis improvisation and elude other cases of improvisation. In addition, our findings
specifically deal with technology use in a limited range of crisis situations: crisis response to
natural disasters. For this reason, generalizability of our findings is likely to be limited.
Further research should assess the adequacy of our findings within other contexts for
improvisation and technology use.

Conclusion

In this work we tackle three aspects of organizational improvisation in crisis response that
have been insufficiently explained so far by the literature. First, we address the paradox of
acting and planning by examining theoretical work on interactions between crisis improvisers.
Second, we clarify ICT support to crisis improvisation by reviewing crisis management
literature. Finally, we inspire from the dynamic capabilities approach to explore the
conceptual ties between improvisation and routine.
Our work identifies five constituents of organizational improvisation in crisis response among
which three main mechanisms that regulate organizational improvisation. We aim at a more
realistic view of crisis response, enriched by the dynamic capability approach. Our results
show that ICT is a potential support for organizational improvisation during crisis. Thus, the
contribution of this work is threefold.
First of all, we rely on the dynamic capability approach to propose that routines are likely to
provide support to improvisation rather than prevent it from developing. We view routines as
providing a referential from which actors can deviate. Furthermore routines are not static and
do not necessarily enclose actors into well-known patterns of actions. As suggested by the
dynamic capabilities approach, organizational members integrate novelty into existing
capabilities thanks to codification, discussion and repetition (Zollo, Winter, 2002).

  172
Secondly we propose some description of the core constituents of organizational
improvisation in crisis response. We make further conceptual propositions about the role
played by these constituents. By doing so, we attempt to respond to the conceptual paradox of
planning and acting that we mention earlier in this article: How can improvisers coordinate
with each other.
Finally, we propose an explanation of the ICT support to organizational improvisation by
highlighting six ICT properties and their support to the constituents of organizational
improvisation. We explicitly state some ICT properties that are generally implicitly taken into
consideration by most studies.

  173
Annex

Table 25. Description of the sample articles

Authors & title Date Journal Name

COMFORT LK, Integrating information technology 1993 Journal of contingencies
into international crisis management and policy. and crisis management
CALLOWAY LJ, and KEEN PGW, Organizing for 1996 Journal of Information
crisis response. Technology
COMFORT LK, SUNGU Y, JOHNSON D, and DUNN 2001 Journal of contingencies
M, Complex systems in crisis: Anticipation and and crisis management
resilience in dynamic environments.
TUROFF M, CHUMER M, VAN DE WALLE B, and 2004 Journal of Information
YAO X, The design of a dynamic emergency response Technology Theory and
management information system. Application
DAWES SS, CRESSWELL AM, and CAHAN BB, 2004 Social Science Computer
Learning from crisis - lessons in human and information Review
infrastructure from the world trade center response.
YUAN Y, and DETLOR B, Intelligent mobile crisis 2005 Communications of the
response systems. ACM

DANTAS A, and SEVILLE E, Organisational issues in 2006 Journal of Contingencies

implementing and information sharing framework: & Crisis Management
Lessons from the matata flooding events in new
zealand.
NTUEN C, BALOGUN O, BOYLE E, and TURNER 2006 Ergonomics
A, Supporting command and control training functions
in the emergency management domain using cognitive
systems engineering.
CARVER L, and TUROFF M, Human-computer 2007 Communications of the
interaction: The human and computer as a team in ACM
emergency management information systems.
MENDONÇA D, JEFFERSON T, and HARRALD J, 2007 Communications of the
Collaborative adhocracies and mix-and-match ACM
technologies in emergency management - using the
emergent interoperability approach to address
unanticipated contingencies during emergency
response.
PALEN L, HILTZ SR, and LIU SB, Online forums 2007 Communications of the
supporting grassroots participation in emergency ACM
preparedness and response.
MCKINNEY EH, Supporting pre-existing teams in 2008 Journal of Information
crisis with it: A preliminary organizational-team Technology Theory and
collaboration framework. Application

  174
  175
 Part 3. The empirical part
The empirical work aims to respond to the limitations of our theoretical findings to research
question II. To do so, we develop an integrative analysis of the 2003 French heat wave crisis
response. This part is structured as follows.
Chapter 3.1 presents the case of the 2003 French heat wave crisis response. We first detail
the specificities of the response network that gathers 23 types of organizations involved in the
crisis response. We then present the specificities of the 2003 French heat wave crisis
response.
Chapter 3.2 details our research design. We present the research objectives and topics of the
empirical work. We then present our methodology, our sources of data and the diverse steps
of our data collection and analysis.
Finally, chapter 3.3 presents our empirical investigation and findings. We respond to
research questions III and IV by providing an integrative framework that emerged from our
coding.

  176
Chapter 3.1. Presentation of the 2003 French heat wave case

Introduction

The heat wave, which occurred during the summer of 2003 in France, had dramatic effects on
the environment and the population. During the months of July and August, there was a
dramatic drought. Lakes and rivers were dried-up enough to block energy supplying
(Salagnac, 2007). The simultaneous effects of the heat and the drought raised the risk of fires
and energy blackouts in electrical and nuclear power plants. Additionally, the 2003 French
heat wave provoked a sudden increase in mortality in the Ile-de-France region, especially
among the elderly. Although the effects of the climate on the French territory and the French
population had already been documented, the health care system failed to anticipate the
increase of mortality caused by the heat wave (Got, 2008). As a result, the death rate was
more than 150% higher than previous years during the two first weeks on the month of
August (Fouillet et al., 2006; Hémon et al., 2003). Figure 38 represents both temperatures and
death numbers in Paris between the 15th of June 2003 and the 15th of September 2003, which
illustrates the simultaneity between increases in deaths and temperatures around the 12th of
August. The response to the heat wave involved a set of 23 types of organizations that we
label the response network.

Figure 38. Death numbers (left ordinate) and Celsius temperatures (right ordinate) in
Paris between 6/25 and 9/15

  177
This chapter describes the heat wave crisis response and is structured as follows. Section 1
presents the 23 types of organizations that compose the response network. Section 2 then
details the response network functioning. Section 3 describes the organizational crises that
were triggered by the heat wave within the response network. Section 4 provides a
chronology of the response to the heat wave crisis.

Section 1. The response network actors

We identified three main groups of organizations within the response network.

First of all, the group of civil protection and police organizations is in charge of assuring
public security and order. In critical situations, these organizations have the additional
responsibility of providing civil protection and emergency services, such as caring for injured
people and organizing rescue operations. This group mainly includes the Paris Firefighter
Brigade (BSPP) 12 and Paris Police services. Both of these services are under the authority of
the Paris Prefecture of Police (PPP). Similarly, the PPP is under the authority of the Minister
of the Interior of France.
Secondly, the group of health care organizations is in charge of taking care of injured and sick
people. This group includes three types of organizations with distinct levels of
responsibilities. Local actors such as emergency medical services and hospitals handle
patients’ flow, diagnosis and care. Regional organizations manage and regulate medical
activities in Paris and its surroundings. Some national organizations from the group of health
care organizations manage the funding and organization of health care activities within
France. Other organizations from this group are in charge of alerting all the health care
organizations when some disaster is likely to happen such as the National Health Watch
Institute (INVS).
Finally, we identified a third group of organizations that were involved in crisis response. We
label this group “outsider organizations” because these organizations were not entitled to
respond to the heat wave. However, outsider organizations had a significant influence on the
development of crisis response and improvisation. Finally, we include patients and citizens in
the response network because some of them were affected by the ongoing improvisation13.
Figure 39 details the connections between the organizations that were involved in the
response to the 2003 French heat wave crisis. Organizations with national responsibilities are
represented on the top. Progressively moving towards the bottom one can find organizations
with regional and then local responsibilities. As the legend explains, unidirectional arrows
signify hierarchical top-down connections. Bidirectional arrows represent strong
cooperational ties. Finally links with no arrow correspond to ad hoc connections that emerged
between organizations during the heat wave. We provide further details about the specificities
of each organizations and each group of organizations in the following paragraphs.

12
All acronyms of organizations that were involved in the crisis response are in French.
13
We did not collect data from victims.
  178
Civil protection & internal
Health care organizations Outsiders
affairs organizations

Internal Affairs Health Ministry Social Affairs

Ministry Ministry

MétéoFrance
COGIC

DHOS DGS Funeral

INVS
companies

AMUF

DDASS ARHIF AP-HP

Town
PPP
Halls

Journalists

Paris
Firefighters SAMU/
PPP local squad (BSPP) SMUR
agencies
SOS
médecins Red Cross
Liberal
EHPAD Hospitals
physicians

Emergency
Rooms (ER)

Intense
cooperation Citizens
ties

Hierarchical
ties

Other
connections

Figure 39. The Ile-de-France Health Network

1.I. Health care organizations

Health services were prominent in 2003, especially in the Parisian Region. According to the
INSEE (the national institute that produces statistics), French health activities represented
more than half of the national budget of France and involved more than 170,000 persons in Ile
de France (e.g. 1/5 of the French employed by the health care industry).14

14
For more detail see www.insee.fr and more specifically
http://www.insee.fr/fr/themes/tableau.asp?reg_id=20&ref_id=sanopte73
  179
In 2003, international reports depicted the French health system as one of the most efficient
health care systems in the world (Le Faou, 2003). However, the French health care system has
been aiming to restrict its budget and rationalizing its processes since the mid nineties (Le
Faou, 2003). Budget restrictions have resulted in actions intended to reduce costs, such as
temporarily closing unoccupied beds and services. For example, the DHOS and ARH
recommend every summer that hospitals close between 10 and 30% of their services. Because
an increasing number of people suffer from long term diseases for which patients obtain fully-
financed treatments, hospitals and emergency units’ budgets are more restricted (Poutout,
2005).
Within the health care system, two main logics coexist: a professional and an administrative
logic (Boudes, Laroche, 2009; Robelet et al., 2005). On the one hand, the health care system’s
mission is to treat patients. This mission governs field actors’ activities and reasoning, in
particular in local organizations such as hospitals, SOS médecins and BSPP. On the other
hand, an administrative logic stems from the necessity to coordinate and regulate health care
activities at regional and national levels (Robelet et al., 2005). The administrative
organization of the response network features this logic. National organizations such as the
DGS and the DHOS make decisions on resource allocation (Le Faou, 2003). National
organizations and regional organizations such as the DDASS and the ARH agencies are in
charge of monitoring health care activities. Some regional organizations such as the AP-HP
directly interact with field actors and are familiar to field actors’ needs. However, their
primary mission is to coordinate health care activities and apply decisions that were made by
national organizations.
Both the professional and the administrative logics influence organizational practices in the
response network. The administrative actors competently make decisions by using statistical
tools, data sets and analysis grids. For this reason, they primarily focused their attention on
risk mitigation objectives when they interacted during the heat wave. Conversely, local
organizational actors had a vivid perception of the new and dramatic aspects of the situation
when the heat wave struck (Milet, 2005). Field actors that include Emergency Room
Technicians (ERTs), firefighters and emergency medical practitioners considered taking care
of patients as the top priority in their work. Some field actors experienced administrative
matters as being constrictive. In the following paragraph, we detail each health care
organization.
1.I.1. National organizations
1.I.1.1. The Health Ministry also called Le ministère de la Santé
The Health Ministry activities are related to health services, health insurance and social
welfare in France. However, this ministry delegates many of its responsibilities to national
organizations such as the DGS (more detail in the following paragraphs). The Health Minister
collaborates with his or her cabinet advisers who play a significant role in decision making in
the ministry.

  180
 1.I.1.2. The National Health Watch Institute or Institut National de
Veille Sanitaire (INVS)
The INVS’s role is to warn and alert when some phenomenon is likely to affect French
people’s health. The INVS’ mission is to watch potential sources of epidemic and injuries at
national and international levels. When the INVS identifies a potential health threat, it then
provides its recommendations to the DGS that implements them. The INVS was created in
1998 but in 2003 most field actors, such as physicians, did not know of its existence and
name.
1.I.1.3. National Health Agency or Direction Générale de la Santé
(DGS)
The National Agency for Health is in charge of developing and implementing programs for
health at the national scale. Coordinating with the INVS to implement the monitoring/alerting
system is part of its responsibilities as well.
1.I.1.4. Hospitalization and Health Care Organizations Agency or
Direction de l’hospitalisation et de l’organisation des soins (DHOS)
This agency’s task is to organize health care activities in France, relying on coordination with
the DDASS and DGS. In particular, this organization is devoted to the allocation of financial
resources. Recently, its mission has expanded to the management of the coordination between
the diverse actors involved in health issues.
1.I.2.Regional organizations
1.I.2.1. Local Agency for Social and Health or Départementale de
l’aide sociale et sanitaire (DDASS)
The DDASS is in charge of social and health matters in French regions and counties. The
DDASS delegates to sub agencies that are located in Paris (DASS 75) and the Ile-de-France
counties (for instance DDASS93, DDASS94).
1.I.2.2. Ile de France Hospitalization Agency or Agence Régionale
Hospitalière d’Ile de France (ARHIF)
The ARHIF manages and coordinates hospitals, in particular when national programs and
projects are implemented. For instance, there has been a specific program to promote hand
washing by professionals. This project provided washing solutions, communication displays,
etc.
1.I.2.3. Parisian Agency for Hospitals Management or Assistance
Publique des Hopitaux de Paris (APHP)
The APHP is in charge of managing 39 hospitals in the Parisian region, which includes all
hospitals in Paris intra muros and some hospitals in the neighborhoods of Paris. The APHP
regulates funding, resource allocation, patient flows and alerting. The APHP hospitals have a
strong identity, calling the head of the APHP the “core”. As a result, conflicts between the

  181
APHP and the DDASS commonly occur because of these two organizations have equal
responsibilities in some hospitals.
1.I.3.Local organizations
1.I.3.1. Hospital personnel
Each hospital is in charge of the quality and the coordination of its health care activities.
Hospital personnel include managers, administrative personnel and medical personnel. While
the managers are responsible for the back office management such as human resources,
information systems, quality and performance measurement, another set of administrative
personnel handle coordination between services and surrounding hospitals and assisted living
facilities. From a medical standpoint, each hospital gathers diverse specialists from
psychiatric to emergency practitioners. Professional values are strong among medical staff
such as physicians, Emergency Room Technicians (ERTs) and nurses. Public and private
hospitals hardly coordinated with each other during the heat wave (Evin, 2004). However,
public hospitals personnel were retrospectively appraised by national organizations for their
courage. Most of the hospitals personnel faced an overwhelming workload and relentlessly
worked at treating patients despite their suffering from the heat and exhaustion (Evin, 2004;
Lalande et al., 2003)
1.I.3.2. Emergency Room (ER) personnel
This is a special group within medical staff. While people who suffer from long-term diseases
get planned treatments, patients who spontaneously suffer from injuries are treated in
emergency services. Emergency rooms were overwhelmed during the heat wave. Officials did
not recognize Emergency Rooms Technicians (ERT) as medical specialists in 2003. For this
reason they were extensively represented by trade unions such as the AMUF.
1.I.3.3. Private physicians
Private medical activity is now very segmented and highly specialized (Poutout, 2005). Some
private physicians practice in double locations: they see patients in a private medical practice
as well as in hospitals and clinics. Coordination between physicians and other health care
organizations is controversial. Some private physicians were criticized for not being available
during crisis response (Lalande et al., 2003). As a result, hospitals were missing staff but
could not contact physicians. During public hearings, the specialization of physicians was
claimed to be a social issue. For older people who have multiple sources of suffering, it is
complicated to find a physician who is experienced and knowledgable about their specific
needs.
1.I.3.4. Assisted living facilities also called Établissement
d’hébergement pour Personnes Âgées Dépendantes (EHPAD)
These types of establishments offer apartments and health support for the elderly. During the
heat wave, administrative actors struggled to grant sufficient access to hydrating resources to
both EHPADs and hospitals. However, EHPAD are eligible for access to resources in
surrounding hospitals in case of important shortages: When the EHPAD staff is overloaded

  182
with patients, their counterparts in hospitals are called upon to take care of them, therefore
alleviating the workload of the EHPAD staff. During the heat wave, mortuary chambers in
many EHPADs were overloaded. In most cases, even the surrounding hospitals could not
manage all the corpses, which provoked more confusion within EHPAD.
1.I.3.5. Physician network for emergency service, called SOS
Médecins
This non-profit organization involves physicians who offer mobile medical intervention for
emergencies at home. As other emergency services, they experienced an overwhelming
increase in requests between the 4th of August and the 14th of August. Public hearings
revealed that they had interacted with journalists since the beginning of the crisis.
1.I.3.6. Emergency ambulance service, called Service d’Assistance
Médicale d’Urgence (SAMU) and Emergency and Ressuscitation
Mobile Service, called Service mobile d’Urgence et de Réanimation
(SMUR)
SAMU and SMUR’s memberships are composed of emergency medical practitioners who
provide mobile and immediate help to people who are injured. It provides phone assistance as
well, with the objective of providing medical advice for emergency situations. Emergency
ambulance services coordinate with hospitals when they cannot treat their patients with
mobile devices. They also coordinate with firefighter squads but provide a different service in
that the SAMU first tries to advice by phone before intervening. Moreover physicians are less
numerous among firefighters than within SAMU. The emergency ambulance services
extensively coordinate with firefighter squads (BSPP). During the crisis, firefighter squads
and the SAMU developed ad hoc information systems together.
1.II. Civil protection & internal affairs organizations
The objectives of the civil protection organizations are twofold. First of all, they have to
maintain public order and protect the civilian population from terrorism and large-scale
accidents. Secondly, they have to provide support to civil protection in critical situations. The
civil protection organizations’ mission is oriented towards security. For this reason, these
organizations abide by strict rules of operation. For instance, the BSPP refused to transmit
data about deaths to the INVS without an official authorization from the Paris Police
Prefecture on the 11th of August.
1.II.1. National organizations
1.II.1.1. Internal Affairs Ministry, called Ministère de l’Intérieur
This ministry is at the head of the civil protection organizations. Its involvement in the crisis
response efforts was initiated on the 12th of August. On the 13th, this ministry intervened by
providing military beds for the purpose of stocking corpses.
1.II.1.2. The Inter-Ministry Coordination Agency, also called Le
centre opérationnel de gestion interministériel des crises (COGIC)

  183
The COGIC organized inter-ministry meetings between the Internal Affairs Ministry, the
Health Ministry and the Social Affairs Ministry on the 12th of August to collectively find
solutions to specific issues such as corpses storage and electrical blackouts.
1.II.2. Local organizations
1.II.2.1. Paris police authorities also known as Préfecture de Police de
Paris (PPP)
This organization is responsible for the protection of citizens. PPP is additionally in charge of
police services and the administrative processes related to identification documents.
1.II.2.2. Paris Police Prefecture local agencies:
PPP local agencies exist in every administrative region, or arrondissement in Paris. PPP local
agencies directly intervene when called. This organization is an equivalent to 911 services in
the United States. During the heat wave, he PPP local agencies collaborated with BSPP and
SAMU/SMUR to assist people who fainted, in particular homeless people.
1.II.2.3. Paris Firefighters Squads also called Brigade des Sapeurs
Pompiers de Paris (BSPP)
One of their missions is to alleviate citizens’ immediate suffering and to provide assistance in
emergency situations. Their area of expertise is not restricted solely to damage from fires due
to the fact that firefighter squads include medical personnel. The BSPP is officially a military
organization, a characteristic that distinguishes it from squads of firefighters from other
regions. Its hierarchy and values are similar to those of the army. On the field, the BSPP
extensively coordinates with the SAMU/SMUR teams in managing patient transfers into
hospitals.
1.II.2.4. Paris Town Hall
The Paris Town Hall provided material support to health care organizations. In addition, they
transmitted alerting information to the public in relation to hyperthermia. Finally, they
collected data about victims when these data could not be obtained from private physicians.
1.III. “Outsider” organizations
These organizations do not extensively coordinate with the other organizations routine
situations. However, they developed ad hoc connections within the response network during
the heat wave.
1.III.1. National organizations
1.III.1.1. Weather forecast agency called in French Météo France
Météo France is in charge of providing national weather forecasts. Prior to the crisis there was
already coordination between Météo France and the InVS. Météo France was in charge of
providing updated information on the weather on a regular basis.
1.III.1.2. The Social Affairs Ministry also called Ministère des
Affaires Sociales et de l'Emploi
  184
This ministry is in charge of social issues such as social exclusion and unemployment. In
2002 as well as on the 26th of May 2003, the Ministry anticipated that the elderly could suffer
from the heat during the summer and sent a fax to local governmental agencies to recommend
extra attention on this specific point. During the heat wave, this ministry was involved in the
response by participating in inter-ministry meetings beginning the 12th of August. However,
the Social Affairs Ministry members remained isolated and stated that they received scarce
information from health care organizations.
1.III.1.3. Funeral homes also called Pompes Funèbres
Funeral home companies usually do not coordinate with hospitals but with town halls and
prefectures. However during the heat wave, the lack of coffins to stock corpses was
problematic. For this reason, hospitals developed extensive coordination with funeral
companies. Furthermore, funeral homes participated in some inter-organizational meetings.
The leading company on the market was the only organization that was able to provide
statistics related to mortality at a quick rate. Given data about coffin orders and the fact that
they corner approximately 25% of the market, it was easy for them to determine an estimated
number of fatalities. Thus they transmitted their data about death rates to ministries since the
12th of August.
1.III.1.4. Journalists
Television journalists from local and national channels as well as local and national
newspapers journalists, such as Le Parisien had frequent interactions with crisis responders.
From the beginning of the crisis, firefighter squads were followed by journalists. News
programs broadcasted interviews with medics and provided advice to avoid hyperthermia to
the public. In addition, some medics participated in television and radio programs to alert the
public to the situation. Journalists heavily criticized the lack of reactivity from national health
care organizations such as the DGS, INVS and the Health Ministry. As the number of
fatalities came as a surprise, journalists investigated how data was managed, which resulted in
a political and social controversy. During the last phase of the crisis response, press
conferences were frequent and some administrative actors permitted inter- organizational
meetings to be recorded as evidence of their willingness to remain transparent. Internet users,
including journalists, covered the evolution of the heat wave and interacted with the
population through articles and comments about web pages.
1.III.1.5. French Emergency Room Technician Trade Union, or
Association des Médecins Urgentistes de France (AMUF)
This trade union was particularly active in July 2003 and alerted the public about bed
shortages. The union leader, Patrick Pelloux, intervened on television by sending out an alert
about the massive amount of deaths on August 10th, asserting that such deaths could have
been avoided and were not natural. On that point he experienced an open conflict with the
DGS director who claimed at that time that he was a liar. Patrick Pelloux and his colleagues
intervened in ad hoc transversal meetings that took place on the 11th and the 13th of August.

  185
 1.III.2. Local organizations
1.III.2.1. Paris Red Cross, called Croix Rouge
The Red Cross is an international non-profit organization. Its activities are diverse, ranging
from commodity collection for poor people to first aid intervention. In France alone, it is
comprised of more than 50000 volunteers. The Paris Red Cross intervened in alleviating the
workload of other crisis responders during the 2003 summer. Its values are oriented towards
charity and volunteering.

Section 2. The response network values

We define the response network as the set of the 23 types of organizations involved in the
response to the organizational crisis caused by the heat wave (more detail on the
organizational crisis in section 3). Our definition of the response network is consistent with
Romelaer’s definition of an organization (2002) in that it gathers individuals and groups that
have regular and stable connections between each other. This enables us to study
organizational improvisation – rather than inter-organizational improvisation – within a
network. In the coming paragraphs, we detail the role of collaboration and hierarchy within
the response network. We then present the actors of the response network.
2.I. Collaboration
We use the term network due to the multiplicity of interactions that took place between the
organizations during the response to the heat wave crisis.
Some organizations had developed close ties of cooperation well before the crisis (Robelet et
al., 2005). For example, health care services to the elderly were officially restructured in 1998
into a network constituted by assisted living facilities, town halls, home hospitalization
services and hospitals. Some other connections between these organizations were similar to a
meshing configuration. For example, the firefighter squads (BSPP) were not only under the
authority of the Prefecture de Police, but also had institutionalized links of cooperation with
emergency medical services such as Emergency rooms, SAMU and SMUR. Generally
speaking, the functioning of the health care system is based on interindividual interaction
(Destais, 2003) and interorganizational connections. Typically, the regulation of the flow of
patients is predefined by procedures but also depends on the informal interactions between
emergency units and friendships between professionals. Finally new links of cooperation
developed between actors from diverse professional groups during the crisis response. For
example, some professionals contacted the health department minister or participated in
institutional meetings.
Collaboration is a widespread value within the network. Each hospital is supposed to
collaborate with local private physicians. The APHP and hospitals also cooperate with the
mobile emergency services by sharing tasks. Similarly, the BSPP handle a significant part of
the emergency medical interventions.

  186
Collaboration implies information sharing. During the heat wave, hospitals provided
information to potential victims in assisted living facilities (EHPAD). Similarly, town halls
were contacted by medical services to collect data about people who were likely to be hurt
from the temperatures, such as old persons. Funeral parlors spontaneously contacted the
Health Ministry and the prefectures as well.
2.II.Hierarchy
The response network does not only include collaboration; Hierarchical values are also
important within this network. For instance top-down and bottom-up coordination coexist
between the Health Ministry and other entities of the network. Hospitals have to validate the
most critical decisions by the Health Ministry and the prefectures, e.g State representative in
local counties. However, the Health Ministry delegates many activities to the DGS and the
DHOS that are responsible for health related issues in France. The Health Ministry also relies
on these organizations for the monitoring of the national health care system. At a more local
level, the DGS and the DHOS delegate some of their prerogatives to the DDASS and the
APHP. Similarly, hierarchy is particularly important between firefighters (BSPP), police
services (the PPP local agencies and the PPP) and the Internal Affairs Ministry.
2.III. Organizational and informational barriers
The concept of complexity is prevalent within the response network. There have been more
than 24 reforms within the French health care system since 1945 (Poutout, 2005), which
explains why some organizations remain unknown within the response network. This was the
case of the agency for health alerts (called Institut de Veille Sanitaire, also known as INVS)
before the 2003 French heat wave.
As Denis, Lamothe and Langley argue, organizations related to health issue classically
involves organizations with divergent objectives and values (2001). It is interesting to note
the example of the Parisian firefighter squads (BSPP), under the authority of police services
and governed by military values (firemen in Ile de France have a military status, on the
contrary to other places in France), that were prone to withhold information based on the
rationale of avoiding public panic. On the contrary, some active medical personnel such as
ERT searched to transmit and publish death rates as much as possible to provoke a reaction
among the public and promote civil mobilization towards the situation of the elderly.
Organizational barriers are important as the network gathers organizations with distinct
missions and values. For example, field actors of the network claimed their opinion was
looked down upon by the administrative organizations because they missed a global view of
the health care system. As an emergency room technician explains:
“The Health ministry advisers were more or less competent. Most of the time they were honest
but their interest narrowed to political and financial objectives. No minister ever had an ear
to listen to professionals, unless they had the same rank” (L’urgentiste, p.26).
Organizational barriers also arise from the gap between professional values and logics within
the network. Health care organizations are oriented towards saving lives where as civil

  187
protection organizations look after keeping calm among citizens. As some health care
institutional actor explained during the public hearings:
“My observation is that hospitals and sanitary institutions are completely disconnected from
the prefect services, may it be right or wrong. I am only making an observation”.
Finally, separate information systems strengthen organizational barriers. Only one federating
information system existed in 2003 to coordinate hospitals and the Health Ministry: the
PMSI. The PMSI is a routine tool used to compile and store data to get an idea of the time
spent by each patient in services. As the PMSI use is restricted to accounting we exclude it
from our analysis. Data sharing was restricted, because IS architectures were segmented from
one hospital to another in spite of the need to develop a new culture of data sharing (Fieschi,
2003). As a firefighter explains during public hearings, the lack of information sharing
prevents organizations from getting a comprehensive view of their own processes.
“The firefighters’ squads have a partial view on of the state of health of the whole population
that is under their protection. We do not receive information about all the deaths, in
particular these that concern people who were at home but who could not call us. We do not
know what happens to victims after their being transferred to hospitals”.
Finally, the frequency of ICT use differs from one organization to another, strengthening
organizational barriers. Field actors, especially physicians, show modest use of emails
(Brooks, Menachemi, 2006) on the contrary to administrative actors. Phones and faxes were
primary means for communication in health care organizations in 2003 (Chen et al., 2010).
This study focuses on four types of technology that were used before and during the
organizational response to the crisis provoked by the 2003 French heat wave: the Internet,
email, faxes and phones. During the 2003 French heat wave, these technologies were helpful.
As parliamentary reports on the heat wave crisis response explain, the Internet, used by
physicians and other actors, helped to obtain information about hyperthermia and death rates.
Actors needed an overview of the situation as well. A director acknowledges in a report:
“At that time, we could not know the exact number of sick people in all the emergency rooms.
We strove to evaluate numbers but, given our means, we hardly managed it”.

Section 3. The 2003 French heat wave as an organizational crisis

In August 2003, the heat questioned the ability of several organizations to perform their
missions, therefore putting these organizations into crises. During the heat wave, civil
protection and health care organizations partially failed at protecting the lives of vulnerable
citizens. Emergency rooms were overwhelmed with patients and the civil protection services
did not manage to prevent outdoor and indoor fatal faints. As an increasing number of people
suffered from the heat, hospital mortuary chambers were overcrowded and funeral homes
were unable to manage the increasing number of corpses. The following week, the Health
Minister, the National Health Watch Institute and the National Health Agency failed to
provide the public with a fair approximation of the number of deaths. It took an entire week

  188
for the Health Minister to enact the emergency plan, also known as the “White Plan”, that
includes directions to cope with a sudden increase of casualties caused by a disaster.
The 2003 French heat wave not only triggered internal organizational crises within the
response network but also paralyzed transversal coordination within the response network. As
a result, the organizations that compose the response network failed at performing their
collective mission: taking care of people and saving lives. More specifically, the extremely
high temperatures led to two fundamental tensions between the health network’s fundamental
objectives and its actual functioning during the heat wave. First of all, hospitals were lacking
information, knowledge, administrative support, energy and material resources that they
usually draw from regional and national administrative organizations. Secondly, the public
health system could not share information and coordinate properly to optimize resource
allocation. While more than 15 employees were mobilized to write one note alerting the
public to the situation, help was desperately needed to simply hydrate patients in hospitals.
Each organization had internally settled crisis management teams and meetings. However,
information was not shared between the groups of civil protection organizations and medical
organizations. For all these reasons, the 2003 French heat wave not only put some
organizations into crises but also triggered an organizational crisis within the response
network.
We argue that the 2003 French heat wave crisis matches our theoretical assumptions about
crises (for more detail about these assumptions, see chapter 2.1). The heat was not the only
source of difficulties that influenced hospitals’ functioning during the 2003 summer. Crisis
was caused by the combination of unexpected triggering events, such as the heat, and inherent
dysfunctions (Milet, 2005). For instance, the overall restriction of the number of available
beds made health care organizations react slower to the sudden increase of casualties. The
Hospitalization and Health Care Organizations Agency (DHOS) ordered hospitals to reduce
their activity in May of 2003. Every summer, the national agencies that are in charge of
managing the health care system and its budget temporarily close around 25% of services in
hospitals. The rationale for this annual closing is that people usually go on vacation and the
occurrence of casualties is likely lower during the summer. In addition, a significant
proportion of medical personnel goes on vacation during the month of August. Thus summer
brings the opportunity to make savings by closing services. Beds are “closed” and stocked,
which limits the hospitals’ capacity to care for large quantities of patients. However, in
August 2003, casualties suddenly increased and the number of patients waiting for treatment
in emergency rooms far exceeded the number of beds available. Opinions of actors involved
in the response network on the bed closings diverge. While some Emergency Room
Technicians (ERT) claimed that the lack of beds jammed the access to emergency rooms and
increased the ERTs’ work overload and stress (Pelloux, 2004), some other reports argue that
there is no correlation between beds closings and the mortality rates (Bastianelli et al., 2003).
Regardless of the divergent opinions on the bed closings, the response network was inherently
dysfunctional because of the multiple misconnections between the organizations included in
the response network. As the Lalande report explains, the administration did not know how to

  189
collect information while the majority of crisis responders who were on the field did not think
about the necessity to communicate with the administration (Lalande et al., 2003). Despite
evidence that some ERTs tried to contact administrations by sending emails or making phone
calls, no one can deny that the response network lacked cohesion:
“Organizational disconnections between health care administrations, between ministries and
between administrations and actors who were on the field restricted possibilities for
information sharing. Every organization that managed the response network nationally [e.g.
INVS, DHOS, Health Ministry, etc] had centralized and developed its own ad hoc information
system to deal with the crisis. However, there was no effective networking between these
organizations. Generally speaking the “center” does not know how to collect information and
the “periphery” does not think of the necessity of communicating information”. (Summary, p.
2)
The response to the 2003 French heat wave crisis was burdened by the lack of information
sharing between crisis responders. In spite of well-established patterns of coordination, most
actors are not used to sharing information. For instance emergency rooms handled extra
activities rather than relying private physicians to complete these tasks. The extra workload
burdened the management of the overwhelming number of patients in hospitals. As some
national reports explains:
“When they were on the field, private physicians and SOS médecins were overwhelmed by
calls and faced some dramatically tough situations. However, numerous crisis responders
such as SAMU personnel, squads of firefighters and emergency staff claimed there was a lack
of coordination between hospitals and other care units. They also mentioned that a massive
number of private physicians who were familiar with the elderly patients, had departed for
their summer vacation. The absence of these physicians created many problems: Medical
personnel was lacking information about the patients’ medical background (…) and were
overwhelmed with requests for death certificates”. (Summary, p. 4).
Our understanding of the 2003 French heat wave is consistent with theories on crisis in two
other aspects: the heat wave provoked both substantial emotional and time pressure. In this
study, we focus our analysis on the crisis coping/response step of the whole episode. Figure
40 represents the combined effects of organizational dysfunctions and heat on the emergence
of organizational crises during the heat wave. Organizational dysfunctions burdened
coordination within the response network while the heat played the role of the triggering
event.

  190
 Figure 40. The 2003 heat wave crisis as an endogenous and an exogenous episode

Section 4. Crisis response chronology

We provide a chronology of the events of the crisis that we developed from the systematic
examination of the interactions between actors. We triangulated this chronology with two
chronologies previously developed in the literature (Boudes, Laroche, 2009; Lagadec, 2004).
4.I. Disparate signals (from the 4th to the 7th)
Loose signal detection started on the 4th. There were disparate signals from emergency
professionals, such as SAMU and BSPP. For example, the DDASS from Britain signaled the
suspicious deaths of three persons under the age of 60. These signals were reported to the
heads of the INVS and the DGS, who expressed their concern regarding the health effects
caused by the heat. One of the Health ministry’s advisers then asked his delegates to write an
alert message to health professionals on the 6th. The writing of the alert message followed a
full validation process that took more than 36 hours. Similarly some of the AP-HP directors
informally met emergency squad members on the 5th to voice their concerns about the heat.
4.II.The beginning of the coping (between the 8th and the 10th)

  191
The situation rapidly worsened. Health care operational actors became overwhelmed and
began to improvise. The AP-HP sent an alert that basically asked the hospital staff to mobilize
itself, help each other with patients and prepare to face difficulties because the weather was to
be hot during the weekend. The AP-HP asked hospitals to act as if the White plan was on the
scene. On Friday morning the situation remained under control. Still, during the afternoon, all
the emergency squads were overwhelmed with patients and requests. During the weekend, the
situation worsened even more for all emergency services: emergency units in hospitals,
SAMU and SMUR, firefighters, SOS médecins. During the weekend many field actors
developed improvisation practices, ranging from the ad hoc information system between
firefighters and SAMU to innovative uses of ice in hospitals. At the end of the weekend,
Patrick Pelloux, the AMUF leader, alerted the public on television and asserted that deaths
were numerous and that the situation was more than critical. Some administrative actors then
participated in the ongoing collective improvisation, such as AP-HP and DHOS, to provide
extra resources.
4.III. What do we do now? (between the 11th and the 13th)
From the 11th, the administrative sphere recognized that the situation was out of control and
attempted to resolve several issues. The catastrophe was taking place and panic emerged
because no organization had a precise idea of the number of deaths. The hospital’s staff was
exhausted, as Patrick Pelloux claimed on television. Transversal meetings between DHOS,
DGS, AP-HP, InVS and Health Ministry advisers took place on the 11th and the 12th.
4.IV. The beginning of the controversy (between the 13th and the 20th)
Since the afternoon of the 13th, the situation got better; the temperatures and the attendance of
emergency services simultaneously decreased. However, the administrative sphere had settled
into heavy data collection processes on the hospitals’ activities, which required hour-per-hour
reports from services in critical situations about patients, deaths and temperatures in rooms. In
the afternoon the Internal Affairs Ministry officially triggered the national emergency plan,
also called “White Plan”. From this time on, the effort to compile statistics of death rates
started and the political controversy surrounding the government’s ability to handle the
situation developed. This was the end of the organizational crisis but the beginning of the
political crisis.

  192
 Conclusion

This chapter presents a chronology of the crisis response to the 2003 heat wave and the
response of network actors. The response network is composed of three groups of
organizations: civil protection and internal affairs organizations, health care organizations and
“outsider” organizations. Undoubtedly, the heat was not the only factor that hindered the
health care and civil protection organizations’ ability to perform its mission: taking care of
people and saving lives. Organizational and informational barriers partially account for the
heat wave crisis, even though collaboration and hierarchy played important roles in the
functioning of the response network.

  193
Chapter 3.2. The research design of the empirical investigation

Introduction

The empirical part of the dissertation consists of a retrospective qualitative study of the
response to the organizational crisis provoked by the heat wave in France in 2003. The 2003
French heat wave took place from the 4th to the 20th of August and allegedly provoked 14,802
deaths in France. In the 2003 summer, temperatures exceeded 105°F (40°C). During the
month of August 2003, the situation worsened with temperatures higher than 115°F during
days and nights. On the 5th of August, surprising deaths were reported. During the whole
month, medical organizations hardly overcame the overwhelming number of incoming
patients. The Health Network struggled to respond to the overwhelming number of patients.
Crisis responders had to improvise collectively to handle the local resources shortages and
uncertainty regarding what could be done to alleviate the patients’ suffering. Crisis responders
developed innovative practices to handle the situation collectively. Some of the crisis
responders extensively used ICT but hardly participate in the ongoing improvisation.
Conversely, evidence of electronic communication between improvisers is scarce.
This empirical investigation addresses the two following research questions:
- Does ICT fit users’ needs with respect to interaction during crisis improvisation? (III)
- How do users benefit (or not) from ICT when they improvise collectively in crisis
response? (IV)
The objective of this chapter is to present the research design of this empirical work.
The objective of our empirical work is to provide an integrative response to research
questions III and IV. To do so we restricted our investigated to three topics (Blaikie, 2007). In
addition, our method follows the principle of the grounded theory.
Our data collection covers three main sources: public hearings, internal archives and external
archives that we selected on the basis of their reliability and their proximity with the research
topics. The public hearings consisted of more than 100 hours of public collective interviewing
completed in the aftermath of the heat wave by members of two investigation commissions.
Internal archives correspond to the documents that the heat wave crisis responders exchanged,
ranging from emails to press releases drafts. The external archives gather essays, academic
works, reports and newspapers’ documentation on the crisis response to the 2003 French heat
wave. In this study we coded emails, faxes and public hearings and completed additional
analysis of external archives. Given the important amount of data we had to analyze, we had
to complete a data reduction by following a theme dictionary (Thiétart, 2007).
Beside it is important to note at this point that we have changed the study research design
since the thesis proposal. We eventually focused our data coding on the content of public
hearings that took place in the aftermath of the heat wave. We attained theoretical and

  194
semantic saturation by the end of the coding of the public hearings.15 Rather than completing
observation and focus groups we relied on external and internal archives to triangulate data.
This chapter is structured as follows. Section 1 defines the study objectives and topics.
Section 2 presents an overview of our methodology choices: a retrospective qualitative case
study. In addition argue our rationale for following the grounded theory (Corbin, Strauss,
1990; Strauss, Corbin, 2008). Section 3 presents our three data sources. Section 4 details the
iterative process that we followed to collect and analyze data.

Section 1. Our research objectives

Our primary objective is to provide an integrative and contextualized response to research

questions III and IV. As a result, we address research questions III and IV simultaneously by
observing interactions that took place during the 2003 French heat wave crisis response.
Our rationale for following this approach is twofold. First, we assume that technology fit,
users practices and organizational context depend on each other. Technology fit depends on
the users tasks and the context of use. Reciprocally, users’ practices around technology have a
moderating effect on user’s needs and technological fit. Second, because reality is made of
complex dynamics (Bourdieu, 1980) and that technology is a core component of our
technology (Latour, 2004; Orlikowski, Iacono, 2001)., there is higher value in an integrative
description than a partial formalization.
Our ambition to complete an integrative investigation has major implications. The scope of
our investigation includes a significant panel of variables, such as crisis responders’
interactions, ICT characteristics and use, improvisational practices. For this reason, we
segmented our investigation into three topics that we detail in this paragraph. We identified
this research technique in Blaikie’s work (2007, p. 6). Our objective is therefore to
simultaneously investigate topics I, II and III:
Topic I. The features of technology use to interact during crisis response.
Topic II. The variables that account for technology use for interacting between crisis
improvisers.
Topic III. The influence of technology use on the interactions that compose organizational
improvisation.
According to Blaikie (2007), empirical investigation covers three topics: “What?”, “How?”,
“Why?”. The “What?” question “requires a descriptive answer” to account for characteristics
of the phenomenon that is under study. In our study, the “What?” question consists of
exploring the nature and dynamics of interactions and technology use during organizational
improvisation in the context of crisis response, which corresponds to topic I. We agree with
Latour (2004) that a rigorous description of concrete life is necessary to avoid the risk of

15
Theoretical saturation refers to the fact that data does not provide any further content. Semantic
saturation occurs when the coding does not make further coding emerge.
  195
creating knowledge that is disconnected from reality or to treat technology as a “black box”.
However, our approach is not only descriptive but also explanatory. Descriptive and
explanatory approaches are interrelated (De Vaus, 2001).
The “Why?” question asks for either the causes or the reasons for the existence of
characteristics and regularities of the phenomenon that is under study (Blaikie, 2007). It
corresponds to the topic II and deals with factors that account for ICT use for interacting
during organizational improvisation in crisis response. Finally, the “How?” question
investigates the effect of the phenomenon under study on organizational life. Accordingly, it
corresponds to the analysis of the influence of ICT use on interactions that compose
organizational improvisation. Figure 41 presents each topic as well as connections between
topics.

Figure 41. Connections between topics

As Figure 41 illustrates, topics are interrelated: The users rationale for using or not using ICT
for crisis improvisation may depend on how ICTs are used by crisis responders. Similarly,
The influence of ICT use on improvisation stems from the features of ICT use. Investigating
all the topics in the same study enables us to provide a global explanation of ICT support to
organizational improvisation during crisis response.

Section 2. Method: An overview

In this section we present the major characteristics of our study. First, our investigation is
retrospective and qualitative. Second, our approach is inspired from the grounded theory
perspective. We also introduce some major implications of the grounded theory approach on
our research design.

  196
 2.I. Our motivation for a qualitative case study
Qualitative methods help to explain phenomena that have not been explored in an extended
manner in the MIS field yet, such as organizational improvisation. Since our definition of
organizational improvisation basically starts from theory, it is necessary to evaluate its
consistency with empirical data, which is another motivation for a qualitative approach.
Our motivation for a qualitative methodology is threefold. First of all, qualitative
methodology is consistent with an exploratory approach. There has been little description of
the constituents of organizational improvisation in crisis response in the literature so far.
More specifically, we know little about how crisis improvisers interact with each other.
Qualitative methodology allows responding to this lack by capturing data “from the inside”
through a process of deep attentiveness (Miles, Huberman, 1994). Secondly, explicating the
nature and dynamics of interactions that compose organizational improvisation is a
preliminary step to study possible relations between organizational improvisation and other
variables. To that sense, a qualitative investigation seemed prior to any quantitative approach.
Thirdly, qualitative methodology is consistent with holistic investigation (Miles, Huberman,
1994).
In this work, we opted for a case study methodology. A case study enables to develop deep
observation of organizational processes and behaviors that compose a specific phenomenon
(Gombault, 2005). It grasps the complexity of organizational improvisation in crisis response.
2.II. A grounded theory approach
In this section, we present the grounded theory approach as defined by Corbin and Strauss
(1990) and its practical implications on the design of our research. We first justify our interest
in the grounded theory approach. Then, we argue that the Straussian perspective on grounded
theory is the most consistent with our pragmatist approach of the research question (for more
detail on our following of pragmatist principles, see chapter 1.3).
2.II.1. Our rationale for a grounded theory approach
Our motivation to use the grounded theory method is threefold. In fact, Locke acknowledges
at least three conditions for following a grounded theory approach (2001). Among them we
retained three main motivations that were further explicated by Sousa and Hendricks (2006):
First, there is deficient theoretical guidance: existing theoretical frameworks are undeveloped
or do not provide satisfying answers to the research questions, which signifies the need for
theory enrichment (i).
Second, Core concepts are loosely coupled to each other or the connections between concepts
are obscure. An inductive approach is adequate to clarify the connections between the
concepts under study (ii).
Third, observation of participants is necessary to respond to the research questions (iii).
Our dissertation project meets these three conditions, which supports our strategy of
following the grounded theory approach.

  197
First of all, the conceptualization of organizational improvisation remains incomplete. In
chapter 2.2 we explain that we lack details on how improvisers interact during crisis response
Secondly, the concepts of crisis and improvisation remain loosely coupled, in spite of
significant academic efforts to clarify the antecedents and consequences of improvisation.
Improvisation occurs in both routine and crisis situations but evidence that improvisation
develops likewise in both cases does not exist. Most researchers argue that improvisation is
adequate to respond to turbulence and uncertainty (Crossan et al., 2005). However,
improvisation does not systematically happen in crisis (Cunha et al., 2003; Roux-Dufort,
Vidaillet, 2003b). As a result, the connections between improvisation and crisis remain
uncertain and vague.
The conceptual links between ICT and improvisation need some clarification as well. As a
result conceptual guidance about technology fit to improvisers is missing, which is a
motivation for our study. We develop this point in previous chapters.
Finally, investigation on improvisation requires the observation of improvisers’ practices to
provide insights about the research question. The theoretical propositions that we make about
the interactions that may compose and regulate organizational improvisation are only
exploratory. Furthermore, as we argue in chapter 1.3, the ontology of improvisation is
inherently practical (Cunha, 2004). Deriving substantive concepts therefore implies
observation of participants’ interactions and actions.
2.II.2. A Straussian perspective
This study adopts the Straussian perspective on grounded theory (1990; 2008). Corbin and
Strauss’s view is one among three major streams in the grounded theory literature. Corbin and
Strauss’s approach differs from Glaser and Strauss (1973) and Charmaz ‘s perspective
(2006)16 in being essentially interactionist and pragmatist.
In Corbin and Strauss view, the world is highly complex and unstable in time and pace:
“We are confronting a universe marked by tremendous fluidity: it won’t and can’t stand still.
It is a universe where fragmentation, splintering and disappearance are the mirror images of
appearance, emergence and coalescence. This is a universe where nothing is strictly
determined” (Strauss, 1993, cited by Strauss, Corbin, 2008, p. 6).
This contingency is the result of complex interactions between multiple factors (see
Assumption 9; p7 in Strauss and Corbin, 2008). Neither actions nor interactions are rational,
which means that means-end analytic schemes are not appropriate according to the grounded
theory perspective. Thus abstraction and concepts get their all significance because they
provide additional insights to professionals.
“Concepts provide ways of talking about and arriving at shared understanding among
professionals. If you don’t have a language, you can’t talk – and if you can’t talk, you can’t

16
Charmaz’s perspective on grounded theory is constructivist. Theory emerges for a collaborative
theory construction process between the interviewer and interviewees.
  198
do, and the basis of many professions is still doing” (Blumer, 1969, cited by Strauss, Corbin,
2008, p. 6).
However, our rationale for following a grounded theory perspective goes beyond ontological
and epistemological consideration. In fact, the Straussian perspective accepts that researchers
develop preconceptions before the beginning of the empirical investigation. On the contrary,
Glaser and Strauss argue that conceptual analysis naturally emerges from empirical data. In
our work, literature exploration was useful to develop research questions, and enhance
sensitivity for field, which corresponds to Corbin and Strauss’s approach (p.37). Finally our
approach is not compatible with Charmaz’ constructivist posture (2006). Even if
interpretation is an important part of the theory building process, we do not intend to co-
construct interpretation with actors. Our rationale is that organizational improvisation and
crisis do exist independently of actor’s perception. For more detail on our assumptions on the
ontology of crises and improvisation, please refer to chapter 1.3.
2.II.3. Methodological implications of our following the grounded theory
approach
We mention here some fundamental principles that stem from the grounded theory approach
and that had major implications on our research design. These principles are presented by
Corbin and Strauss (Corbin, Strauss, 1990):
Even if we present data collection and data analysis separately, the two processes are
completed simultaneously. According to the grounded theory approach, data collection and
data analysis are interrelated processes (Corbin, Strauss, 1990). Data analysis starts as soon as
first bunches of data are collected. Moreover data analysis also influences data collection, as it
is suggested by the principle of theoretical sampling. We therefore collected data purposefully
from places and informants that supported the development of our findings. For this reason,
we adjusted our instrumentation strategy on the outcome of the microanalysis and the coding
processes.
Second, we follow an iterative process for theory building that goes back and forth between
data and theoretical propositions and that is described by Figure 42. The analysis process
appears to be generally inductive. Still, the coding process is guided by abduction of
hypotheses (Haig, 1995, 2005; Koenig, 1993) and a deduction from confrontation with rival
hypotheses that are based on theory (Orton, 1997).
Third, our analysis is concept centric. It articulates connections and comparisons between
concepts rather than data. Even if concepts are grounded in data, we aim at analyzing
relationships between concepts rather than data (Glaser, Strauss, 1973; Strauss, Corbin,
2008).

  199
Figure 42. The iterative process for theory building (From Bandeira-de-Mello, Garreau,
2008)

Open Coding Axial Coding Selective coding

THEORETICAL THEORETICAL THEORETICAL

SAMPLING SAMPLING SAMPLING

MicroAnalysis
CATEGORY HYPOTHESIS THEORY
VALIDATION VALIDATION REFINING
THEORETICAL
SAMPLING
DATA

Section 3. Instrumentation

Instrumentation refers to the techniques that researchers use to collect data (Miles, Huberman,
1994). Data collection on improvisation is challenging. Moorman and Miner characterize
improvisation by an almost restricted delay between action and decision and provide a scale
of organizational improvisation (1998a). Even though this scale is useful in questionnaires,
improvisation is difficult to detect for an external observer who mainly perceives the
extemporaneous dimension of improvisation (Ciborra, 1996b). For Weick (Weick, 1988),
observing crisis improvisation is all the more complex as sensemaking is retrospective in
critical contexts. Finally, the beginning and the end of improvisation are difficult to grasp in
that there is sometimes no firm sign from actors that they are going to improvise.
We rely on four techniques to collect data about improvisation. First of all, improvisation is
easier to observe retrospectively even though the retrospective biases can threat the validity of
our findings (for more detail on validity please refer to chapter 3.3). Secondly, we focus
observation on tangible elements of organizational improvisation, e.g. interactions between
improvisers and facts to optimize data validity. Thirdly we rely on theoretical sampling
principles to optimize our chances to observe crisis improvisation. Finally, we rely on the
reality check techniques to compare empirical evidence of improvisation with the definitions
of improvisation that we found in the literature.
  200
 3.I. Retrospective observation
In this study we retrospectively observed interactions between the crisis responders of the
2003 French heat wave. For this reason, we had to manage specific biases related to our
retrospective research design.
The first challenge of this retrospective study is to make account of past practices as faithfully
as possible (Weick, 1993). Before collecting data, we had anticipated some potential
difficulties related to data collection and analysis. Even though the public hearings were
completed in the weeks that followed the heat wave, some potential biases can threat data
validity. Table 26 details some of these biases.
First of all, the public hearings offer retrospective statements from the crisis responders. For
this reason, we suspected that the participants of the public hearings had a tendency to
transform past event because of a retrospective bias, a hindsight bias and some emotional
bias. Due to their similarity to interviews, one could think that public hearings restricted to
consensual subjects. In the case of the 2003 heat wave, national political and social
controversies took place. For instance, French families were accused of not taking care of the
elderly (Le Grand-Sébille, Véga, 2005; Richard, 2008). In addition, the public criticized the
government for its handling epidemic deaths during the heat wave. Due to these
controversies, the participants of the public hearings were questioned about their
responsibilities and their ability to complete their job at the time of the heat wave. To
conclude, taboo issues and lack of trust were scarce during public hearings.
We had to manage our personal biases as well: preconceptions, temporal biases and
conceptual misleadings. There has been an extended discussion in the literature on the risk of
biases that are specific to retrospective studies. To some researchers, “the use of retrospective
accounts in management research needs to be seriously questioned” (Golden, 1992). For this
reason, we had thought about the possible means to avoid its impairing effects on the quality
of our work.
To face these difficulties, we extensively relied on triangulation the diverse sources of data.
We first lead triangulation between the contents of public hearing sessions. We then
compared internal archives and public hearings. Finally, the external archives were used to
triangulate data from public hearings as well. For instance, we had to check that we did not
overestimate the frequency of email use in 2003. We also rely on Klein and Myers’s
guidelines to manage our potential biases (1999). In particular, the principles of dialogical
reasoning and suspicion call for sensitivity to possible inconsistencies and distortions within
findings. Finally, the grounded theory approach advocates sensitivity to potential rival
hypotheses in the theory building process (Glaser, Strauss, 1973).
We provide here two illustrations of how the triangulation process enabled us to understand
interactions. The first example illustrates a taboo and responsibility issue. The Paris Police
Prefecture’s director account was nuanced by two of the BSPP officers. The former claimed
that the Prefecture discovered for the first time the gravity of the situation and the number of
deaths on the 12th of August:

  201
“It is thanks to the Paris mortuary autopsy service that we understood that the situation was
serious. (…) But we were already on the 12th”.
The BSPP officers later mentioned during the public hearings that their superior called the
Prefecture to know whether they were authorized to provide information about the death rates
to journalists on the 8th of August. Some journalists supported the BSPP officers’ claim about
this interaction. By triangulating these two claims, we understood that some interactions
between the BSPP and the PPP had been eluded by the PPP director because they revealed
some dysfunctions within the PPP.
Similarly, our analysis of the external archives supported triangulation of data between public
hearings session. For instance, the Evin report contradicts the AP-HP director claiming that
no patient had been abandoned, on the basis of some ERT and firefighters’ claims (Evin,
2004, p. 93).
We provide a second illustration of the triangulation process between public hearings and
archives. While the Health Ministry advisers claimed that they had no idea of the gravity of
the situation before the 11th of August. The emails that were exchanged and received on the
7th and the 8th of August reveal that one of these advisors had recommended that the DGS
issue an emergency message about the effect of hyperthermia. This interaction was evidenced
in external archives as well. First, the Evin Report mentions that the AP-HP called one of the
Health Ministry advisers on the 7th of August (2004, p. 97). Furthermore, Patrick Pelloux
depicted in his essay his preliminary warnings to the Health Minister at the beginning of the
month of August (2004).

  202
 Table 26. Major biases in our analysis

The type of Definition

methodological issue
Retrospective bias Tendency to embellish or transform past events or actions to
maintain self-esteem (Nisbett, Wilson, 1977).
Hindsight bias Some significant inconsistencies in the respondent’s discourse
(Fischhoff, Beyth, 1975).
Emotional bias Tendency to transform the reality because of emotion (Mitchell,
Thompson, 1994).
Taboo & responsibility Tendency to avoid topics or transform one’s statements because of
issues the negative image of the organization they may convey.
Respondents may lie or over-rationalize past events and actions
(Grawitz, 2001)
Lack of trust issues Tendency to retain information that may jeopardize the informant
status if revealed.
Preconceptions from Preconceptions about the context of the crisis response (Grawitz,
the researcher 2001)
Temporal bias from the Tendency to translate one’s context to the past (Marrou, 1954).
researcher
Conceptual misleading The concepts under study do not refer to any reality for the
respondents and are pure intellectual construction (Harris, 2001).
3.II.Observing ICT and improvisation through interactions
Analyzing ICT support to organizational improvisation requires opening the “black box” of
technology and the analysis of the ICT artefact as embedded in social practices. Orlikowski
(1996), Vaast and Washalm (2005)’s represent some of the prominent authors that have
investigated IS from the practice-based view.
Many authors advocate a practical approach of the ICT artifact in empirical studies. After all,
the raison d’être of ICTs is to respond to user’s daily needs. As a result, the ICT artifact
embbeds in everyday practices (Orlikowski, 2007). Moreover, due to the lack of definitions
for ICT, studies misrepresent technology as unproblematic in itself (Orlikowski, Iacono,
2001). On the contrary to that, the ICT artifact, defined by “those bundles of material and
cultural properties packages in some socially recognizable from such as hardware and/or
software” by Orlikowski and Iacono (2001), intervenes into organizational processes
(Masino, Zamarian, 2003). For example, emails are not only recognized through their
functionalities. They also represent a communication mode and relate to new practices for
organizing work and coordinating from distant locations (Panteli, Fineman, 2005).
But how can we concretely observe practices? And how can we plan for a strategy to collect
data from a practical approach? There is a developing debate on the relevance of the notion of
practice for scientific observation. To a certain extent, every action is a practice and it is
difficult to discern what is a practice from what is not (Pesqueux, 2009). Suchman, Blomberg,
Orr and Trigg (1999) respond to this issue by suggesting the practical view of IT is possible

  203
through observations of daily work practices and interactions. Partly for this reason, we
approach ICT use through interindividual interactions, our unit of analysis in this study. For
more detail on our focus on interactions, please refer to chapter 1.2
3.III. Theoretical Sampling
Theoretical sampling concerns multiple stages of our empirical investigation. Generally
speaking, theoretical sampling is a method that aims at collecting data from places, people
and events that will maximize opportunities to develop concepts (Corbin, Strauss, 1990;
Strauss, Corbin, 2008).
We use theoretical sampling in two steps: in the predefinition of the research design that
precedes data collection and during data collection.
First, we deduced from our conceptual work the case we wanted to study, following Mason’s
recommendations (Mason, 2002). This effort goes beyond choosing a field case, but also
consists of setting the boundaries of the case that are the level of analysis and the unit of
analysis (Miles, Huberman, 1994). This led us to the following strategy. We first focused our
investigation on the case of the 2003 French Heat Wave. We then studied organizational
improvisation within the response network that was composed by organizations involved in
the crisis response.
First of all, we focused on the 2003 French Heat Wave because it has been formally called a
“crisis”. By reviewing literature, we also noticed improvisation was a major dimension of
emergency practices. Improvisation is perceived as a necessity by some authors from the
health care management field. Physicians sometimes have to improvise by diagnosing
(Kirmayer, 1994). Given time pressure, uncertainty and the need for personal, decision
making in health care organizations is complex (Hunink, Glasziou, 2001) and requires
improvisation to respond to unanticipated scenarios (Anderson, McDaniel, 2000).
We focused our investigation on crisis response in the Ile-de-France because the situation was
particularly dramatic in this part of France: buildings were not adapted to the weather, many
families and employees were gone for vacation which made the crisis all the more difficult to
cope with. Moreover, the high density of population implied a more intense attendance to
hospitals during the heat wave.
Preliminary reading of external archives confirmed that our choice to study the 2003 French
heat matches to our research questions: There is evidence that actors had to be surprisingly
innovative and adjust each other to emergent coordination mechanisms. For example, nurses’
creativity was a major condition for continuity of care: they reused painting hooks to hang
hydrating tincture. Responsibilities for ad hoc data gathering and transmission emerged
among hospitals managers.
We then used theoretical sampling to iteratively select segments of data, as recommended by
the grounded theory authors (Strauss, Corbin, 2008). We refer to this practice as data
reduction in this chapter. We used a theme dictionary to iteratively select data for coding and
analysis.

  204
 3.IV. Reality check
We got inspired from Harris’s work on the word courage (2001) so as to check what we
define improvisation definitely ring a bell in respondents’ minds. To do that, we compare her
or the description of improvisation in empirical data is consistent with theoretical definitions
of improvisation. We deduced the reality check grid from our reviewing and comparing of the
definitions of improvisation in the management literature.
By doing so, we have identified five dimensions of improvisation. In other words,
improvisation consists of a i) new and ii) creative iii) resource rearranging, in a iv) short delay
between decision and action, that permits v) adaptation. We then considered every action that
endorsed these five dimensions as improvisation. In Annexes, Table 34 accounts for the
dimensions of improvisation in existing definitions.

Section 4. Data sources

We detail each of the three main sources of data.

First of all, we relied on the content of public hearings as primary data in our analysis. The
Public hearings took place in the aftermath of the heat wave and aimed to understand the
response network failures. The first round of public hearings was completed in and involved
59 crisis responders. Crisis responders were selected to participate depending on their roles
and their interactions during the crisis response. Each public hearing involved the commission
committee and one or several crisis responders. We focused on hearings by people who
experienced directly the heat wave. The commission used a snowball sampling technique to
contact people. They also relied on previously published reports to contact the participants of
the public hearings.
Secondly, we coded some of the internal archives that were joined by the participants of the
public hearings: Emails, internal reports, letters, faxes, press releases. We used some internal
archives as secondary data as well.
Thirdly, our investigation included the analysis of a large panel of external archives as
secondary data. In the aftermaths of the heat wave, a significant amount of investigators
searched for some explanation for the delayed enacting of emergency plans. External archives
are composed of national and international reports, academic work and books on the heat
wave. External archives enable us to understand the economic and social context of the 2003
French heat. In addition, it provides insights on the French health care system and the
professional values of the groups of actors that compose the response network. Archives do
not only help to understand the basic functioning of the health care system but also to identify
major events, that is to say the beginning and end of the crisis response.
Table 27 presents the three main sources of data that we collect in this study. In this table we
specify the themes that appear in each source of data.

  205
 Table 27. Data sources

Data sources Type of data Use of data source Related themes (see the theme
dictionary)
Public Primary data Identification of Interactions configuration, settings
hearings interactions and and development.
actors Organizational context: crisis
Coding settings, structure, organizational
improvisation, organizational
functioning.
Individual characteristics.
Technological characteristics.
Internal Primary data Identification of Interactions development
archives: interactions Events and chronology
faxes, emails, Coding
press releases
External Secondary data Triangulation of Events.
archives: events Economic and social context.
Essays, Chronology of Organizational structures
academic events and ICT infrastructure
articles, interactions
reports, Understanding the
newspaper organizational and
articles environmental
context
4.I. Public hearings
Public hearings articulated into two rounds. The first round of public hearings took place in
September 2003 during the completion of the Jacquat report. The verbal content of these
public hearings was totally transcribed in 134 pages. The whole content of these public
hearings is available online17. The second round of public hearings took place in February
2004 during the completion of the Evin report. Similarly to the first round, the verbal content
of these public hearings were fully transcribed into 575 pages and is online.18 Each session of
the public hearings involved several actors. First of all, the members of parliament who were
completing the reports asked questions to one or several crisis responders. For this reason,
public hearings are somewhat similar to conversations between several interlocutors.
Our analysis focused on public hearings that involved crisis responders. We excluded from
analysis the content that did not relate to the dictionary themes, such as experts’ statements
about the future developments of the national health watching system. Table 28 presents the
numbers of the actors from the response network that participated in the public hearings as
well as the number of hearings that we analyzed. Our analysis focused on the public hearings
of 52 crisis responders in 36 hearing sessions. Each hearing session lasted at least 60 minutes.

17
http://www.assemblee-nationale.fr/12/rap-info/i1091-t1-3.asp#TopOfPage
18
http://www.assemblee-nationale.fr/12/rap-enq/r1455-t2.asp
  206
 Table 28.Organizational affiliation of the public hearings participants

Organization Number of Number of

participants hearings
Paris Police Prefecture 1 1
Paris Firefighter squads (BSPP) 2 1
Health Ministry 4 5
Social Affairs Ministry 2 2
DHOS 2 2
AP-HP 2 2
INVS 1 2
Météo France 2 1
DGS 2 3
AMUHF 2 2
SAMU 1 1
Internal Affairs Ministry 4 2
EHPAD 4 2
Paris Town Hall 1 1
Journalists 1 1
SOS Médecins 4 1
Hospitals 8 3
ARH 2 1
Red Cross 2 1
Liberal physicians 4 1
Inter Ministry Cooperation 1 1
Total 52 36
Public hearings helped to identify the major actors of the crisis response. In addition, they
provided data on the organizational context, interactions, individual and technology
characteristics.
Our rationale for analyzing public hearings rather than leading retrospective interviews is
threefold. First of all, retrospective biases were likely to be stronger in relation to
retrospective interviewing than public hearings. The heat wave occurred 6 years before our
empirical investigation. For instance, crisis responders were likely to embellish past
experiences or even get confused about the order of events. Secondly, public hearings took
place in the aftermath of crisis response and captured details about ICT use. On the contrary
retrospective interviews implied a great effort for respondents to remember about their
interactions during the heat wave. Thirdly, the content of public hearings fits the research
design in that they focused on interactions and information transmission.
4.II.Internal archives
Two types of documents compose the internal archives. The first type of documents – emails,
faxes and letters – describes the information exchanges between the crisis responders during
the 2003 heat wave. These documents were made public by their senders, which means that
we had no grant to have access to all the emails that were exchanged during the 2003 heat

  207
wave crisis response. The second type of document corresponds to charts, press releases and
copies of internal reports that were completed during the heat wave. We focused our analysis
on all the emails and faxes that were joined to the public hearings. We used some documents
of the second type as secondary data, such as the description of the functioning of the
emergency medical services that was joined to public hearings.
We coded 199 email and 58 faxes in a similar manner as public hearings. We coded the
content, as well as the document header, e.g. organizational logo, title or object of the letter,
signature, etc. We then relied on the categories that emerged from the analysis of email and
faxes to refine and test our model. For instance, we assessed whether these could be found in
all sources of data.
We also relied on email and faxes to triangulate the chronology of events and interactions.
However, existing academic work and reports present exhaustive chronologies of events that
occurred during the 2003 heat wave (Boudes, Laroche, 2009; Lagadec, 2005; Lalande et al.,
2003), so we simply checked that the internal documents illustrated the established
chronologies of the heat wave crisis response. Table 29 reports the number of emails that we
analyzed.

Table 29. Number of analyzed emails and faxes

Date Number of Number of

analyzed analyzed
emails faxes
6th of August 1 5
7th of August 2 5
8th of August 16 3
9th of August 0 0
10th of August 0 0
11th of August 22 3
12th of August 17 15
13th of August 52 16
14th of August 79 6
15th & 16th of 20 5
August
Total 199 58
4.III. External archives
External archives include essays, national and international reports and academic papers.
These documents present structured analyses of the implications of the 2003 heat wave.
Because of dramatic fallouts, the 2003 heat wave has attracted many researchers and experts
so far. As a result, the crisis response to the 2003 French heat wave has been significantly
documented. We included 20 newspapers articles in the external archives in spite of a low
reliability (Milet, 2005).

  208
We view external archives as a major source of secondary data in our research (Thiétart,
2007). First of all, we use them to establish the chronology of events during the heat wave. In
addition, we used these documents to triangulate data. Finally, we used external archives to
refine our coding and our model. External archives refer to four types of documents: i) Essays
represented in Table 24, ii) academic work represented in Table 25, iii) reports represented in
Table 26, and finally iv) newspaper articles, represented in Table 27.
The search for external archives mainly went through Internet access and informal requests to
respondents. Analyzing archives and documents we face three main issues: i) How to grant
the archives’ content is true? ii) When analyzing, how to avoid biases from the researcher? iii)
How should we treat data that has already been treated by someone else? To cope with these
issues, we got inspired from historical methods of analysis of documents and archives.
The first issue corresponds to a well-known issue in history. Indeed there is little or no
evidence that archives objectively describe the past. Analyzing archives, one rather
investigates how the archive’s author transcribed and interpreted events than objective reality
(Marrou, 1954). The second issue refers to the difficulty for the researcher not to extrapolate
from his or her experience of the present to understand the past.
The third issue refers to the integration of secondary data into empirical research. Secondary
data refers to “sources of data that has been collected by others, not specifically for the
research question at hand” (Frankfort-Nachmias, Nachmias, 1996). Secondary data can be
found in archives, e.g reports, newspapers articles, academic papers and internal documents,
but have already been selected and treated accordingly to a specific research design. As a
result, they do not perfectly match with the requirements of our work. Meanwhile, the
archives usually represent an important source of secondary data and includes a wide
diversity of format and content, which facilitates triangulation (Cowton, 1998).
Given the input archives bring to this work, we assessed their reliability. Data reliability is an
important criterion to assess the quality of an academic work (Corbin, Strauss, 1990). In
particular, sampling validity “is concerned with any sampling that occurred in the selection of
the texts to be examined (…)” (Harris, 2001). To assess sampling validity of the external we
collected, we classified them. This assessment enabled us to exclude the less relevant archives
for analysis. Documents differ according to two criteria.
The first criterion was proximity between the content of the document and the core of
investigation. Obviously, improvisation, interactions and CT are the core concepts of our
investigation. Still, qualitative investigation is holistic (Miles, Huberman, 1994) and requires
analytic ties between core concepts and broader variables (Strauss, Corbin, 2008). In terms of
proximity, we therefore distinguish core, connected and remote contents. In our search for
archives document, we retained core archives on the crisis response, the context of the 2003
French Heat wave and on the health care system functioning. Archives also differed in terms
of scientific reliability: some are scientific studies and their content may therefore be
considered as secondary data (Thiétart, 2007). Others are not supported by any methodology
and rather provide testimony and opinions. The scientific reliability of archives is somewhat

  209
related to their format. For example, research articles rely on approved methodology. Official
reports are likely to offer less rigorous methodology but they gather a rich content: emails,
internal notes, public interviews from actors.

Table 30. Classification of essays

Title, authors & source Proximity Reliability

Les systèmes de santé en questions: Allemagne, France, Medium Medium
Royaume-Uni, États-Unis et Canada, A-L. Le Faou., 2003, Paris,
Éditions Ellipses.
Canicules, La santé publique en question, L. Abenhaim, 2003 High Medium
Editions Fayard
Urgentiste, P. Pelloux, 2004, Editions Fayard Medium Low
Les chiffres de la canicule. Lettre d’information de Pénombre, High Low
2003, December
Heat Wave: A social autopsy of disaster in Chicago, E. Medium High
Klinenberg lu. The university of Chicago Press

  210
Table 31. Classification of academic work

Authors, Title & Source Proximity Reliability

Impact sanitaire de la vague de chaleur d’Août 2003 : Premier High Medium
résultats et travaux à mener. Bulletin épidémiologique
Hebdomadaire – Novembre 2003
The heat wave in France in August 2003, D. Hémon, E. Jougla. Medium High
Revue épidémiologique de santé publique, Février 2004
Média et Santé publique : l’exemple de la canicule pendant l’été Medium Medium
2003 en France, Boyer, L., Robitail, S., Debensasion, D.,
Pasquier, P. Revue d’épidémiologie et de santé publique, 2005
Understanding the French 2003 Heat Wave Experience: Beyond High High
the heat, a multi-layered challenge, P. Lagadec. Journal of
contingencies and crisis management, 2004
August 2003: Reflections in a French summer Disaster, Thirion, High High
Debeusasorn. Journal of crisis and contingencies management,
2004
Boudes T, and Laroche H (2009) Taking off the heat: Narrative High High
sensemaking in post-crisis inquiry reports. Organization Studies
377-396.

  211
Table 32. Classification of reports

Authors, Title & Source Proximity Reliability

Rapport de la mission d’expertise et d’évaluation du système de High Medium
santé pendant la canicule 2003, Lalande, Legrain, Valleron,
Meyniel, Fourcade. Ministère de la santé, de la famille et des
personnes handicapées, 2003, 61 p.
Surmortalité liée à la canicule d’Août 2003. Rapport d’étape – High Medium
Hermon, Jougla. INSERM, septembre 2003. 59 p.
Apport Information Canicule : La crise sanitaire et sociale High Medium
déclenchée par la canicule, D. Jacquat, 580 p.
Impact sanitaire de la vague de chaleur d’Août 2003 en France, Medium Medium
M. Ledrans, H. Isnard. Institut de veille sanitaire (INVS),
Octobre 2003, 120 p.
Continuité et permanence des soins libéraux Durant l’été 2003, High Medium
J-C Cuenat, C. Daniel, R. Olivier, T. Roquel. Inspection
générale des affaires sociales, 63 p.
Rapport sur les rapports d’enquête sur la canicule, Lagadec, High High
Schepens, Laroche. Revue pluridisciplinaire en sciences
humaines, 275 p, 2004
La France et les Français face à la canicule : les leçons d’une High Medium
crise, Létard, Flandre, Lepeltier, 2004, 405 p.
Rapport n° 1455 au nom de la commission d’enquête sur les High Medium
conséquences sanitaires et sociales de la canicule, F. Aubert, 573
p.
WHO Regional Committee for Europe (2003) Heatwaves: Medium Medium
impacts and responses, WHO, 11p.

  212
 Table 33. Clasification of newspapers article

Source, Title & Date Proximity Reliability

Le Monde, La sécheresse persistante fait craindre un été High Low
catastrophique, 20th of July 2003.
Le Monde, Canicule et sécheresse, un double choc exceptionnel, High Low
6th of August 2003.
Le Monde, La canicule s'installe pour une semaine, 6th of High Low
August
Le Monde, La canicule a provoqué lundi la mort d'un homme de High Low
32 ans dans le Sud-Ouest, 8th of August
Le Monde, Jean-Pierre Raffarin et Roselyne Bachelot n'ont pas High Low
créé de cellule spéciale pour gérer la crise, 10th of August
Le Monde, La vague de chaleur devrait durer encore huit jours, High Low
7th of August 2003
Le Monde, La canicule serait responsable de la mort d'une High Low
cinquantaine de personnes en Ile-de-France, 12th of August
2003
Le Monde, les urgentistes s'alarment du nombre de décès, 12th High Low
of August.
Le Monde, La prévention, remède essentiel contre High Low
l'hyperthermie, 13th of August
Le Monde, Nonchalance, 13th of August High Low
Le Monde, hospitaliers s'estiment à présent confrontés à une « High Low
épidémie de coups de chaleur », 13th of August
Le Monde, Canicule : premières réactions du gouvernement, High Low
13th of August
Le Monde, Douze jours pour prendre conscience d'un « drame High Low
humain », 15th of August
Le Monde, Le Plan blanc est activé, 15th of August High Low
Le Monde, Les autorités sanitaires envisagent 3 000 morts liées High Low
à la chaleur, 15th of August
Le Monde, LUCIEN ABENHAÏM, directeur général de la santé High Low
« Cela correspond à ce que l'on observe dans les épidémies de
grippe », 15th of August
Le Monde, Mourir, en France, sous la canicule, 15th of August High Low
France caught cold by Heatwave, Bulletin of the World Health High Low
Organization 2003, 81 (10)
Libération, comment les chiffres ont flambé, 18th of December High Low
We preliminary read these documents to refine our theme dictionary. We then systematically
classified the content of these archives. We used Document Summary Form to analyze
external archives, as recommended by Miles and Huberman (1994). We analyzed 20
newspaper articles and 26 other documents that we present in the coming tables. Figure 43
and Figure 44 presented in the Annex are examples of our 46 document summary forms. We
structured the document summary form on the basis of the themes that compose the theme

  213
dictionary (for more detail on the theme dictionary, please refer to final section of this
chapter): crisis, technology, improvisation and interactions. We added some extra memos to
analyze the longest documents.

Section 5. Data handling

In this section we detail how we handled the important amount of data in this investigation in
our analysis. To complete data reduction, we relied on a similar technique as the one used to
elaborate the interviewer guide: we elaborated a theme dictionary (Thiétart, 2007).
5.I. An iterative process
We iteratively collected, read and coded data to optimize our chances to focus on the most
relevant data. After settling our research objectives and methods, we started collecting data.
We then started to build the theme dictionary for data reduction, e.g. the exclusion from
analysis of the data that did not relate to our research questions. However, our first coding
experience led us to add or delete data in the theme dictionary. Figure 43 describes data
collection and analysis as an iterative process.
In the first step, we collected the first external archives such as national and international
reports. We extensively relied on literature and chapter 2.3 to build the preliminary version of
the theme dictionary, presented in Table 35. We then modified the theme dictionary in an
abductive manner.
We made a preliminary reading of the public hearings and some external archives. On the
basis of our readings we refined the theme dictionary. For instance, we made multiple
additions and deletions of subthemes in relation to ICT. The first version of the theme
dictionary contained the ICT properties that we had deduced in chapter 2.3. However, we did
not recognize these properties in public hearings and external archives during the first weeks
of analysis. We then responded to this gap between our theme dictionary and data in two
manners. We included additional ICT features from literature reviewing in the dictionary
themes. In parallel, we focused on interactions to investigate ICT use. To do so, we included
subthemes on interactions in the dictionary themes. Eventually, we reidentified some ICT
properties in the final steps of our analysis.
Similarly, the preliminary reading of external archives led us to include trust in the dictionary
theme because the national reports questioned the INVS and DGS members’ reliability in
transmitting information. However, the reading of public hearings led us to nuance our
understanding of this criticism because the INVS and DGS members claimed their integrity.
Emails supported their claim. Consequently, we reformulated the theme of trust as reactivity
and induced its definition from data.
In the third step, we analyzed external archives and then included cultural and professional
values in the theme dictionary. Since that time, we did not have to include additional theme in
the theme dictionary.

  214
In the fourth step, we coded data in public hearings and internal archives, leading to a
preliminary version of the model. The coding of public hearings enabled us to attain
theoretical saturation, which led us to exclude interviews from our data collection techniques.
In addition, we understood at that time the necessity to expand data collection to emails and
faxes to get more specific data on interactions. Once we attained theoretical saturation within
the whole set of data, we extensively discussed the empirical model with colleagues and
senior researchers, which helped to refine our findings.
Finally we reanalyzed external archives on the basis of our model and tested rival hypotheses.
For more detail on this process, please refer to chapter 3.3.

Figure 43. Data collection and analysis as an iterative process

5.II. Data reduction

We completed data reduction to focus the coding of primary data and the analysis of
secondary data. Data reduction helped us to avoid getting lost into the diversity of data. The

  215
data sources do not restrict to data about organizational improvisation and ICT use by the
crisis responders during the 2003 French heat wave. For instance, a significant part of the
public hearings data contains information about the possible improvements of the health
watching system. However, a certain amount of information contained in data sources does
not directly relate to our research questions. Finally, we used a theme dictionary to specify the
meanings of the themes that we analyzed, which contributed to clarity of thought.
Due to large amounts of data, we consider that data reduction is consistent with the grounded
theory approach. In addition we use here the conditional/consequential matrix defined by
Strauss and Corbin (2008, p. 94) to represent the importance of the contextual variables in our
analysis.
Figure 44 represents these variables. First of all, macroeconomic variables partially account
for the medical personnel improvising during the heat wave. For example, crisis responders
had to tackle shortages in beds due to budget restrictions in the French health care system.
The functioning of the health network as well as organizational structures account for
misunderstandings that occurred between medical personnel and other organizations during
the crisis response. At this point, it is important to note that excluded the political aspects of
the crisis response. Undoubtedly, political rivalry played a role in the crisis responders’
behaviors during the 2003 French heat wave. The heat wave resulted in a political fiasco that
led the Health Minister and the DGS director to quit (Boyer et al., 2005). Furthermore,
conflicts between crisis responders emerged partly because of political divergences (Pelloux,
2004). However, we exclude the political aspects of the 2003 French heat wave in our
analysis. Political conflicts aroused in the latter part of the crisis response and are loosely
related to improvisation (for more detail see the crisis response chronology in chapter 3.1).
We selected data for analysis on the basis of four themes. The theme dictionary tables can be
found in the Annexes.
First, we tried to collect data on the interactions that took place during the crisis response.
This type of data includes subthemes such as interactions configuration, interactions settings
and interactions development and is presented in Table 36.
Second, we investigated various aspects of organizational crisis. We identified most of these
themes in our reading of external archives. We then reviewed their definition in the literature
on crisis response. They are presented in Table 37.
Third, we included organizational improvisation components in the theme dictionary. They
can be found in Table 38.
Fourth, we investigated organizational structures and organizational functioning as
contextual variables. They are presented in Table 39.
Finally, we collect data on individual characteristics and ICT properties that are likely to
influence the interaction dynamics and therefore organizational improvisation. These themes
are presented in Table 40 and Table 41.

  216
For each type of data, we provide its definition and some of its possible values. The theme
dictionary was originally driven from our literature review on the concepts of improvisation,
crisis, ICT characteristics and ICT use. However, we have refined the theme dictionaries
through preliminary data collection and analysis.

Figure 44. The conditional matrix

  217
 Conclusion

This chapter describes our research design for investigating organizational improvisation and
ICT use during the 2003 French heat wave. Our research objective is to provide an integrative
answer to research questions III and IV. We theoretically sampled our case study. Following a
grounded theory approach, we theoretically sampled our data. We have iteratively completed
data collection and data analysis. By doing so, we have focused our data collection on three
sources of data and have structured our analysis around four themes.

  218
 Annex 1. Reality check tool

Table 34. The five dimensions of improvisation

Definitions of Dimensions of improvisation

improvisation Temporal Unexpected Creativity Resource Adaptation
contradiction novelty rearranging
“A process of X
introducing changes to
an initial model while
maintaining continuity
in the performance”
(Preston, 1987)
“Thinking and action X
emerge
simultaneously and on
the spur of the
moment” (Ciborra,
1996a)
“Convergence X X
between acting and
planning, and
reworking of pre-
composed material in
relation to
unanticipated ideas”
(Vendelø, 2009)
“Organizational X
improvisation refers to
the convergence of
conception and
execution” (Moorman,
Miner, 1998b)
“Composing and X X
performing
extemporaneously,
conception as action
unfolds” (Berliner,
1994, cited by
Bansler, Havn, 2003)
“Unexpected X X
practices, deviation
from routines,
puzzling or random
behaviors calling for
interpretation”
(Lanzara, 1999)

  219
“Action is taken in X
spontaneous and
intuitive fashion”
(Crossan, 1998)
“Action that fills the X X X
gap between routine
organizational
procedures and events
in the course of daily
work” (Suchman,
2007)
“To cope or X X
ingeniously adapt to a
set of circumstances”
(Preston, 1991, cited
by Vera, Crossan,
2005)
“Decision as action X
unfolds” (Moorman,
Miner, 1998a)
“Devising resourceful X X
solutions to intractable
problems”. (Meyer,
1998).
“The conception of X X X
action as it unfolds by
an organization and/or
its members drawing
on available material,
cognitive, affective
and social resources”
(Cunha et al., 1999c)
“Improvisation is the X X X
ability to recombine
chunks of past
experience into new
patterns of action”
(Rerup, 2001)

  220
Annex 2. Document summary forms for secondary data

Figure 45. First example of document summary forms

  221
 Figure 46. Second example of document summary forms

  222
 Annex 3. Dictionary themes

Table 35. Preliminary version of the theme dictionary

Themes Subthemes and their definitions

Structure and context Coordination mechanism (Mintzberg, 1979)
Economic situation: Any macroeconomic variable that significantly
impacted crisis response
Individual characteristics Age, Gender, Hierarchical status.
ICT properties (see chapter Graphical representation: visual representation of data thanks to colours,
2.3) shapes and photos.
Modularity: the ability to add or delete some applications
Data centralization: the ability to stock the whole data on a single server
Data calculation: the computerized treatment of data
Many-to-many communication: the ability to communicate with multiple
persons simultaneously
Virtuality: the degree to which data can be copied and transmitted with no
material support
ICT use ICT use frequency: the degree of use of technology in crisis response
Technology perception: the degree to which users perceive technology as
a support or a constraint (Orlikowski, 1992)
Crisis conditions of Rigidity threat: the degree to which actors do or do not adapt their
emergence and crisis behaviors to cope with the situation (Staw et al., 1981)
experience Crisis experience: the degree to which the actors perceived the situation as
critical (Billings et al., 1980)
Communication problems: the impact of information centralization and
retention on crisis response (Smart, Vertinsky, 1977)
Preparation to crisis Crisis planning in organizations: the degree of anticipation and
response preparation to critical situations (Lanzara, 1983; Smith, 1990)
Organizational Deviation from established use of resources
improvisation (see chapter Expertise leadership
2.3) Boundary spanning
Minimal structures

  223
 Table 36. Interaction theme dictionary
Subtheme
The number of interacting
persons
Participants organizational
origin
Interaction role
Social and friend ties and
their strength
Interaction start and end
Interaction objective
Interaction novelty
Interaction theme
Interaction regulation coding
Relational
Development
Narration development
Interaction channel
ICT use frequency for
interactions
Degree of alignment on the
interlocutor
ICT use experience
 
Definition
Determines the number of persons interacting within one interaction.
Depending on the number of participants, the interaction
development and outcome may differ
The professional groups or community of practice(Wenger, 1999) to
which each participant belongs.
The role that is endorsed by each participant during the interaction
(Turner, 1988)
“The combination of the amount of time, the emotional intensity, and
the reciprocal services which characterize the tie” (Granovetter, 1973;
Lincoln, Jon, 1979)
Determines the beginning and the end of an interaction
Participants’ objectives or motivation to start an interaction. It may
also reveal some strategy (see individual factors). Even if participants
did not have the initiative to interact, they may have followed
particular procedures that are rationalized. (Turner, 1988)
The extent to which participants already had the same type of
interaction and what differs from interactions participants have been
used to. Also describes the extent to which participants were not
expecting the interaction to occur (see spontaneity)
The subject on which people interact. It is the subject of the
informational content for the interaction
The rules, patterns that regulate interactions (Goffman, 1970)
Time that is available to develop trust and complicity during a
particular interaction (McGrath, 1991)
The extent which previous later interactions recount the relatedness
of earlier interactions (Rafaeli, Sudweeks, 1997),
The communication means that are used to interact
The frequency of use of ICT by a participant.
The extent to which the participant aligns one’s own use of words,
communication practices and ICT use depending on interlocutors
Determines the experience of use of participants and the status of ICT
as a tool for work and communication
224
 Table 37. Organizational crisis theme dictionary

Subtheme Definition
Complexity The extent to which the outcome of the interaction or the interaction itself
depends on other interactions or events.
Uncertainty The extent to which the present situation was unanticipated (defines the
concept of surprise that characterizes crisis) and the extent to which the future
is difficult to anticipate., (in particular when it comes to the outcome of the
interaction)
Triggering event The first signal that starts the crisis response.
Emotional Pressure The extent to which the situation is stressing participants.
Time pressure The extent to which cannot take their time to interact
Contradiction The situation is perceived as contradictory for the participants
Rigidity threats Describes the organization’s tendency to urge and strengthen preplanned
procedures in reaction to crisis (staw et al., 1981). It may questions the
organizational ability to improvise or, on the contrary, stimulated
improvisation
Centralization Describes the organization’s tendency not to delegate and centralize power
(Smart, Vertinsky, 1977)
Entropy Defines a disequilibrium in terms of energy that individuals spent to achieve
their tasks (Katz, Kahn, 1966; Morin, 1977)

  225
 Table 38. Organizational improvisation theme dictionary

Subtheme Definition
Deviation from
established patterns
of action The extent to which organizational members deviate from routine and plans
Spontaneity The ability to individually or collectively to elaborate something new in
response to a particular situation (George, Jones, 1997).
Extemporaneity The extent to which an individual or a collective action is done without
preparation

Table 39. Organizational structures and functioning theme dictionary

Subtheme Definition
Hierarchy Describes the formal distribution of power and responsibilities within
organization
Organization mission Defines the primary organizational raison d’être.
Organizational culture Describes the major values that exist in regards to work, workers and
organizational functioning (Martin, 1992)
Technological Describes the hardware resources, software resources and competencies
infrastructure (Byrd, Turner, 2000)
Location The geographical locations of an organization.
Interorganizational Determines entities the organization connects to (both in routine and
Connections crisis situations)
Sub systems Describes the sub parts of the organization.
Communities of practice Defines the professional group that gathers individuals having common
interests and practices (Wenger, 1989)

  226
 Table 40. Individual characteristics theme dictionary

Sub-Theme Definition
Reactivity Describes the personal ability to react to events or to provide help to co-
workers.
Professional Values Personal expectations in regard to professional activities and professional
and expectations values in general. Depends on the individual’s status in the organization.
Reference group The set of people the participant perceives as belonging to his or her work
environment that defines the social world of work in which he or she engages
(Lawrence, 2006)
Crisis experience The personal or professional experience of crisis and emergency situations.
Crisis response
participation The willingness or the ability to participate into crisis response.
Collaboration Describes the personal experience of collaboration and values in respect to
cooperation
Cooperation Describes the personal experience with previous collaboration and personal
values that are related to cooperation.
Personal experience The extent to which the participant suffered from Heat Wave outside his or her
of the heat wave professional activities. Accounts for a deep emotional experience of heat wave

  227
 Table 41. ICT properties theme dictionary

Sub-Theme Definition
Flexibility in Defines the technology ability to operate well in many different environment;
functionality e.g input quality and quantity (Knoll, Jarvenpaa, 1994)
Flexibility in use Defines the technology ability to provide different informational output and
use (Knoll, Jarvenpaa, 1994)
Flexibility in Defines the technology ability for rapid adjustments with minimal efforts
modification (Knoll, Jarvenpaa, 1994)
Calculation Whether the actors rely on data selection and complex treatment for crisis
response
Data Centralization Affects the ability to access different data from different sources through a
single interface
Virtuality Refers to the non material characteristic of data that implies a facilitated
recovery, replication and exchange. Conversely, materiality refers to the use of
additional objects for communication
Reliability Describes the extent to which participants can rely on telecommunication
devices for work
Technological Defines strengths and weaknesses of technologies with respect to possibilities
affordance they offer the people that might use them (Gaver, 1991)
Information load Corresponds to the amount of information that is received and treated by the
participant in a specified delay
Many-to-many
communication Defines reciprocal information sharing among more than two users
Relation Intimacy Refers to the strength of virtual communication (Calloway, Keen, 1996)
Relation The number of individuals with whom a person interacts (Calloway, Keen,
multiplexity 1996)

  228
Chapter 3.3 ICT, communicative genres, organizational improvisation
and emptiness during the 2003 French heat wave crisis response

Introduction

This paper applies the grounded theory method to understand how ICT supported the
organizational improvisation that took place during the response to the 2003 French heat
wave crisis.
In August 2003, the weather in France suddenly became scorching, attaining record
temperatures (46 C, e.g.130 F) for the first time since 1976. August is usually a time for
vacation and the population enjoyed the heat during the first few days. But the reality of the
heat wave quickly became less glamorous. The number of health complications from the heat
rapidly increased; the elderly, especially in urban areas, were not able to combat hyperthermia
and 14,802 people died in France between the 4th and the 20th of August. The heat wave was
depicted as a natural disaster that provoked an organizational crisis in hospitals (Lalande et
al., 2003). Not only were health care units overwhelmed with people suffering from heat
related illnesses but they also faced unexpected shortages of medical supplies, water, ice, air
conditioners and even nurses. Hospital morgues were overcrowded with corpses and many
electrical devices overheated and broke down. Moreover, physicians struggled to diagnose
and cure hyperthermia because medical knowledge on heat waves was almost nonexistent in
France. The Health Ministry19 did not finish collecting information related to the number of
death. Because of this lack of information, the Health Minister and his advisers did not
understand the gravity of the situation and therefore delayed the enacting of emergency
guidelines such as the White Plan.
The crisis response revealed several different examples of improvisation. Nurses
spontaneously substituted missing perfusion equipment by wall hooks to hang up hydrating
solutions. Hospitals were lacking air-conditioned rooms. To cool down several places at one
time with small effort, nurses motorized the available air conditioners by placing them on
wheels. Firemen and some emergency organizations invented a system of code to differentiate
between deaths caused by the heat from those other resulting from other causes. Finally, some
administrative actors rented refrigerated vans used to transport food to manage the large
quantity of corpses that were overcrowding morgues.
During the 2003 heat wave crisis response, improvisation progressively became
organizational but could not grow to the largest scale possible, e.g. all the groups of actors
who were involved in the crisis response. For instance, most administrative actors, in spite of
having access to crucial resources, failed to participate in the ongoing improvisation.
Interestingly enough, many improvisers had face-to-face meetings and relied on phones to
interact. On the contrary, groups of actors who improvised little or not at all extensively used
emails and faxes. Finally, the actors who used technology to coordinate during organizational

19
The definitions of acronyms are provided in chapter 3.1
  229
improvisation represent a very small proportion of the overall quantity of the crisis
responders.
This study aims to bring some clarity to this mysterious contrast between ICT use and
organizational improvisation. This requires understanding why technology supported, but
from some reason, prevented improvisation from fully developing into an organizational
form. More specifically, our work addresses the two following questions:
- Does ICT fit users’ needs with respect to organizational improvisation during crisis
response?
- How do users benefit (or not) from ICT when they interact to improvise collectively in
crisis response?
These two questions respectively correspond to research questions III and IV in the
dissertation. Theories on the fit between technology, tasks and contingence seem to be the
most appropriate to analyze this phenomenon. However, we argue in this paper that a theory
building approach is required to fully understand the role that ICTs played in the 2003 French
heat wave crisis response. Taking this view into account, we apply the grounded theory
method to build a theoretical framework. Our findings suggest that actors needed to tackle
what we call organizational emptiness during crisis response. To respond to this emptiness,
crisis responders had no choice but to improvise. However, our analysis reveals that the crisis
responders’ participation in the ongoing improvisation depended on their communicative
genres, e.g. the communication practices and strategies that are widely recognized and applied
within a group of actors (Orlikowski, Yates, 1994; Yates, Orlikowski, 1992). When they
improvised, the field actors communicated in a “fervent” genre, characterized by personal
involvement and emotional interactions. Conversely, a “dispassionate” communicative genre
was commonplace between administrative actors. The dispassionate genre consisted of
protecting oneself from potential criticism stressing one’s transparency in electronic
communication. By preventing the administrative actors to mention and take part in
improvisation, the dispassionate genre burdened the development of organizational
improvisation.
The presentation of our research is structured as follows. After a succinct presentation of the
concepts of crisis and improvisation, we assess the relevance of the theories of fit to evaluate
the extent of technological support during crisis improvisation. Then, we justify our interest in
the grounded theory method Finally we present a comprehensive view of our findings and
discuss their theoretical and managerial implications. Annexes provide more detail about the
process of coding.

Section 1. Theoretical background on improvisation and crisis response

1.I. Organizational crisis during disasters

Disasters refer to massive material and human damage (Perry, Quarantelli, 2005) that disturbs
organizational functioning enough to trigger organizational crises (Pearson, Claire, 1998).

  230
This is what happened during the 2003 French heat wave (Lalande et al., 2003). The
temperature elevation was a natural disaster: it favored drought, fires and electricity breakout
all over Europe. It also triggered organizational crisis within the French health network by
threatening its functioning and primary mission: taking care of citizens (Lalande et al., 2003).
Organizational crisis is a critical experience that threatens organizations’ primary goals and
values (Hermann, 1963). Because of important constraints, organizational crisis requires
quick and innovative response by improvising (Crossan, 1998). Stakes and costs of a
mishandled crisis are not only material but also social (Dynes, Quarantelli, 1976), which
causes important stress and emotional pressure (Milburn et al., 1983a 1983; Smart, Vertinsky,
1977).
This means that organizations have to manage both disaster response and potential
organizational crises during disasters. They participate in alleviating material and human
damage but also cope with unexpected incidents that put them into crisis. For instance, during
the 9/11 or Katrina disasters, hospitals, emergency units, and police services struggled to save
lives. They also spent time and energy reestablishing the electrical infrastructure that was
necessary for the functioning of equipment (Mendonça, Wallace, 2007). During the heat
wave, the response network responded to the heat but also needed to manage their own
shortages of resources, lack of information and overwhelmed nurses (for more detail on the
response network, please refer to chapter 3.1).
Organizations face important complexity and uncertainty while responding to organizational
crisis in disasters. Uncertainty refers to the organizations’ difficulties to forecast events
related to the crisis. Uncertainty requires the organization to be resilient by developing not
only anticipatory but also improvisational skills (Rerup, 2001). Complexity stems from the
intricate set of interdependent outcomes from the crisis (Milburn et al., 1983a). Improvisation
requires organizations to make do with available resources and to be responsive enough to
quickly develop new solutions by improvisation.
1.II.Organizational improvisation for crisis response
Improvisation has been frequently associated with critical situations (Ciborra, 1996a;
Hutchins, 1991; Rerup, 2001) but is not restricted to situations during which time and
emotional pressure are important (Ciborra, 1996a). Improvisation refers to a creatively acting
on the spur of the moment. It implies a limited delay between acting and planning (Moorman,
Miner, 1998b), which results in a seemingly extemporaneous action (Weick, 1998).
Improvisation includes novelty of action, also called bricolage (Cunha et al., 1999c) that
consists of deviating from established uses of resources.
Improvisation frequently occurs during crisis response due to strong time pressure,
complexity and uncertainty (Crossan et al., 2005; Cunha, 2004). Improvising can be more
relevant than planning during crisis response because it enables adaptation to the specificities
of the situation (Rerup, 2001; Waugh, Streib, 2006). However, improvisation can become
difficult to manage since it involves various individuals and groups (Hutchins, 1991). It
requires efficient interactions (Weick, 1998) and the ability to quickly understand one

  231
another’s actions. Examples of failed improvisation provide evidence of how misguided
interactions can lead to misunderstandings of intentions and threaten coordination. The Mann
Gulch disaster, analyzed by Weick (1993) and Ciborra (1996b), is a vivid illustration of the
influence interactions have on coordination when improvisation occurs. As Weick explains
(1993), more people could have survived the fire if the demonstration of improvisation had
been perceived as a way to escape fire rather than suicide.
“Dodge [one of the survivors] yelled at the crew to drop their tools, and then, to everyone’s
astonishment, lit a fire in front of them and ordered them to lie down in the area it had
burned. No one did, and they all ran for the ridge” (p. 629).
Listening and adjusting to individual initiatives is of primary importance when leading
collective improvisation. Taking this into account, how can actors be more vigilant to each
other when they have less time to adjust to the ongoing action and experience important
emotional pressure? Some resources are crucial to meet this requirement. For instance,
Moorman and Miner highlight the importance of organizational memory to rely on common
structures (Moorman, Miner, 1998b).
This study explores ICTs as other resources for interaction during improvisation. By doing so,
we propose an alternate view of technology and improvisation. While most studies present
improvisation as a natural way to handle technology in use (Ciborra, 1996a; Ciborra et al.,
1999), development (McGann, Lyytinen, 2008) and implementation (Heeks, 2006), this study
focuses on ICTs as resources for improvisation during crisis response (Mendonça, Wallace,
2007), that we label “crisis improvisation”.
1.III. Technological fit to crisis improvisation
In this section, we address the first research question, entitled “Does ICT fit users’ needs with
respect to organizational improvisation during crisis response?”
To respond to this question, we first discuss the relevance of the theories of fit as a potential
theoretical background to investigate ICT support to crisis improvisation.
So far, academics have not provided clear evidence of whether technology significantly
supports crisis improvisation or not. First of all, it appears that the increasing use of ICTs has
made some tasks for crisis response more feasible. Without the large panel of tools we have
access to nowadays, there is no doubt that crisis management would be trickier. For example,
Web 2.0 not only enables information gathering and representation, but also contributes to
social ties and collective intelligence (For-mukwai, 2010; Vieweg et al., 2008). Geographical
Information Systems (GIS) enable locating and tracking crisis responders. More basic tools
such as email, forums and Computer Supported Collaborative Work (CSCW) tools enable
collaboration, particularly with the use of maps. For instance Google Maps was extensively
used during the swine flu pandemic to mark and share knowledge about victims all over the
world20. Google maps supported improvisation by allowing users to update the

20
http://maps.google.com/maps/ms?ie=UTF8&hl=en&t=p&msa=0&msid=106484775090296685271.0
004681a37b713f6b5950&ll=32.639375,-110.390625&spn=15.738151,25.488281&z=5
  232
representation of the pandemic evolution. Finally, ICTs enable crisis responders to
communicate almost instantaneously thanks to quick and easy information transmission. By
doing so, ICTs respond to the crisis responders’ need for quick information feedback (Dynes,
Quarantelli, 1976; Quarantelli, 2007).
However, researchers have provided enough evidence to question whether the ICTs meet
crisis responders’ needs when improvisation is needed. Technology is not always user-
friendly in that its design restricts the users’ actions to predefined processes and parameters.
Beyond reliability issues (Jaeger et al., 2007), functionalities may even obstruct crisis
response given users practices (Dawes et al., 2004). For example, replying to message and
managing ICT devices compels users to take time that could otherwise be used for action
(Jaeger et al., 2007). Similarly, if actors are too dependent on technology, loss of data during
a crisis situation can bewilder users enough to jeopardize crisis improvisation (Hutchins,
1991). In other words, technology may threaten improvisers’ creativity if not managed
properly (Mendonça et al., 2002).
The crisis management literature equally presents pros and cons about ICT use for crisis
response. Due to the lack of comparison between the pros and the cons, the crisis
management literature fails at specifically defining the degree to which ICTs fit to crisis
responder’s needs when they have to improvise. One can consider responding to this lack by
relying on the theories of fit.
The theories of fit have been widely used in the IS literature to examine ICT support to
organizational processes. In spirit, these theories assume that a minimal level of coherence
between resources, processes and structures is necessary to attain performance (Drazin, Van
de Ven, 1985). In the IS literature, the technology fit theories take into account technology
characteristics, tasks characteristics, individual and contextual characteristics. Table 42
represents some widely accepted theories of technology fit in the IS literature and their
constructs (inspired from Zigurs and Khazanchi’s work (2008)). For instance, the media
richness theories compare the specificities of the ICT users’ tasks with the media
characteristics. Knoll and Jarvenpaa assess whether ICT can meet the users’ changing needs
in turbulent environments (Knoll, Jarvenpaa, 1994). Finally, Task Technology Fit and IT
appropriateness theories include the perception of an ICT by its users into the evaluation of
technology fit.

  233
 Table 42. Theories of fit in the IS literature

Name of the theory Analyzed constructs

Media Richness Theory (Daft et al., Media richness, task equivocality, task
1987; Dennis, Kinney, 1998; Rice, uncertainty
1992)
Technology fit in highly turbulent Environmental turbulence, flexibility in
environment (Knoll, Jarvenpaa, 1994) use, flexibility in modification, flexibility
in function
Task Technology Fit (Goodhue, Task complexity, task interdependence,
Thompson, 1995; Zigurs, Buckland, task uncertainty, individual use of
1998; Zigurs, Khazanchi, 2008) technology
IT appropriateness (Khazanchi, 2005) Respondent position, organizational size,
perceived appropriateness of an ICT.
In this study, we argue that the theories of technology fit cannot be applied to the context of
crisis improvisation without some conceptual enrichment.
These theories miss details regarding the specificities of improvisation. While improvisation
can be characterized as a complex and uncertain task (Zigurs, Buckland, 1998), it is also
extemporaneous and situated (Ciborra, Willcocks, 2006). However, the theories of fit do not
take into account these fundamental features of improvisation. In other words, one cannot
anticipate an improviser’s identity, e.g. who will improvise and the improvisers’ practices, e.g.
what improvisation will consist of. Even though these aspects are likely to play a role in ICT
use during improvisation (Brigham, Introna, 2006), they cannot be defined prior to an
improvisation, which prevents researchers to provide an overall assessment of ICT fit to
improvisation.
Based upon our research, there is a relative similarity between individuals’ experience of
technology when examined through the theories of fit. However, crisis response involves
numerous groups of individuals as well as various ICT tools and tasks. For instance, the crisis
response to Katrina involved companies, emergency units, administration and so on. Each
organization had access to various ICTs such as the Internet, phones, PDAs (Chua, 2007;
Davis, 2008, December; Ives, Junglas, 2006a; Quarantelli, 2006). Tasks were diverse as well,
ranging from treating patients to managing extra transportation systems. Thus the ICT
services that enable interactions between crisis responders are likely to not fit every individual
involved in crisis response.
Finally, the concept of fit implies a conceptual distinction between context, tasks and
technology (Zigurs, Khazanchi, 2008). However, crisis improvisers can improvise on
technology. In this case, crisis responders adapt to the technological tools they have at hand,
even though these tools are not adequate to the users’ tasks (Konsynski, Tiwana, 2004).
However, the theories of fit do not take into account the users’ adaptation to technology.
Against this background, we believe that observing and analyzing field data would provide
additional insights to our research questions. As we present in chapter 3.1, the need for a
  234
theory building approach to the question of ICTs and crisis improvisation exists21, regardless
of the limitations of the theories of fit. To respond to this need, we rely on a research design
based on the grounded theory.

Section 2. Methodology and coding

Chapter 3.1 of the dissertation provides more details on the 2003 French heat wave crisis
response. More specifically, this chapter describes the field actors and the administrative
actors of the response network. Chapter 3.2 details our research design, in particular our data
collecting and analyzing. The methodology and coding section details the coding process that
contributed to the emergence of the concepts or organizational emptiness, dispassionate
communicative genre and fervent communicative genre.
The research method in this study is qualitative and retrospective. We followed the grounded
theory approach, as defined by Corbin and Strauss (Corbin, Strauss, 1990; Orlikowski, 1996;
Orlikowski, Hofman, 1997). The grounded theory approach aims to build alternate
explanations of social phenomena by rigorous coding processes. Following a grounded theory
methodology, we identified categories from data. We then transformed them into concepts,
which correspond to the abstraction process as defined by Strauss and Corbin (2008).
Our analysis was iterative. By going back and forth between data and findings, we formulated
hypotheses from the data and we confronted these hypotheses with more data. We coded data
through the four steps that Corbin and Strauss recommend to follow (1990): microanalysis,
open coding, axial coding. We describe each of these steps.
In the very beginning of our analysis, we completed a microscopic analysis of each word of
the segments of the public hearings transcripts that we had selected for analysis. Several
categories emerged from this analysis, as represented by Step 1 in Figure 47. We present
some examples of the categories that emerged from our microscopic analysis in Table 43.
“Difficulty to see and understand” referred to the lack of understanding of the situation for
decision makers. For instance, the administrative actors lacked information about the number
of victims of the heat wave. Due to this lack, they experienced difficulties in settling clear
directions to field actors. On their side, field actors had vivid experiences of heat and death.
Additionally, they received little help from administrative actors and lacked resources. For
this reason, they felt isolated, which refers to the category “Difficulty to be listened to”. The
category “Lack of communication” referred to the misconnections between groups of actors
within the response network. Finally, the code “death” refers to the real time experience of
physical suffering, isolation, emotional exhaustion and confusion.

21
For more detail on our motivation for following the grounded theory approach please refer to
chapter 3.2
  235
 Table 43. Some examples of categories

Name of the Verbatim Explanation of the category

category
Difficulty to see and “We had no clue of The decision makers hardly make sense of
understand what was going on” the situation or they misinterpret the
situation.
Difficulty to be “She told that we The actor has the feeling that his or her
listened to would manage the superior do not care about his or her
situation anyway” difficulties in coping with the crisis
Death “People were dying. Physical experience of death and exhaustion
Nurses were by field actors.
exhausted”
Lack of resources “Patients were sitting Lack of beds, medicines, air conditioner, etc
on chairs, desperately
waiting for treatment”
Lack of “I do not know who I The crisis responders who are supposed to
communication am supposed to act collectively do not contact each other.
contact anyway”
We extensively relied on memos to further analyze these categories and transform them into
codes. For instance, we reflected on the category “difficulty to see and understand” and we
noticed that the administrative actors of the response network experienced an overall
confusion during the heat wave. Some of them had a medical background and for this reason
they felt concerned about the temperatures and its effects on patients’ health. Despite this
concern, most of the administrative actors did not receive any alarming message in the
beginning of the crisis.
Since the 11th of August, the administrative actors realized that patients were dying. Since that
moment, some of the administrative actors experienced an existential crisis when they faced
their own inability to react properly to the situation. Finally they indirectly experienced death
when the media started to mention that more than 5,000 people had died between the 4th of
the 10th of August. From this analysis, we extended the category “death” to the indirect
experience of death and suffering.
We then relied on the memos we had written during our open coding to complete the axial
coding. The axial coding consists of refining and clarifying the codes by comparing and
discussing the conceptual ties between the subcategories (Corbin, Strauss, 1990). The axial
coding started in January 2010 and finish in August 2010. We first identified “difficulty to see
and understand”, “difficulty to be listened to” and “death” as the subcategories of the code
entitled death. Then we related the notion of death and the categories “lack of resources” and
“lack of means”. This process corresponds to Step 2 in Figure 47.
Thanks to discussion with colleagues, we challenged the assumption that the administrative
actors were experiencing death metaphorically. Instead we suggest they experienced some
loss of control over their work coupled with the lack of resources, which provoked an overall

  236
confusion about the response network’s ability to fulfill its mission. Thus, we refined the code
“death” as the code “organizational hole” that referred to the crisis responders’ experience of
lacking something. We understood then that experiencing an “organizational hole” led the
some crisis responders – mostly the field actors - to improvise. Conversely, the administrative
actors experienced organizational holes but did not improvise. This process is represented in
Figure 47 as steps 3 and 4.
By detailing the concept of organizational hole, we realized that the heat wave crisis
responders experienced diverse lacks concomitantly. Therefore, we had to find a more
appropriate term to convey the idea that crisis responders did not experience one specific lack
at a time but rather experienced something more complex. To clarify this point, we made
some semantic research on the distinction between organizational holes, vacuum or void. The
term hole conveys the image of an identifiable lack. However, our analysis rather suggested
that lacks were interdependent. To us the notions of vacuum and void were inadequate
because they do not convey the idea that people had a positive reaction to the lacks by
improvising. These terms defined a “space filled with nothing”, therefore conveying the idea
of non-existence. On the contrary, our analysis suggested that the lack of resource or means
made crisis responders identify opportunities for ad hoc solutions. In our view the notion of
emptiness was relevant because it defined a space that is “vacant or unoccupied” and
therefore free for reinvention. In old English emptiness referred to the lack of occupation. For
instance young people would refer to their emptiness to signal they were single (Thesaurus
dictionary). Therefore the notion of organizational emptiness refers to the fact that some
resources or means were interdependently missing which prevented actors to complete
processes on the usual track. This final step of the axial coding is represented as Step 5 in
Figure 47.
Finally, we completed the selective coding by clarifying the logical connections between the
codes. More specifically, we defined the influence of organizational improvisation on
organizational emptiness and the other concepts that compose our framework. This final step
brought the opportunity to test the relevance of each code and to refine our framework. We
presented the framework in multiple seminars, which helped us to challenge the overall
coherence of the framework. For instance, we had initially made a distinction between crisis
responders’ interactions and the dispassionate communicative genres. This conceptualization
was challenged when some colleagues asked us the difference between an interaction and a
communicative genre. Further reflecting on this comment led us to delete the code
“interaction” from our findings. We then realized that two communicative genres were
coexisting into the response network.

  237
 Figure 47. The coding process steps

Section 3. Findings

We present here the framework and the concepts that emanated from our data analysis.
Our framework shows that organizational emptiness challenged crisis responders and
compelled them to improvise together. However, crisis responders’ participation in
organizational improvisation depended on their communicative genres. A communicative
genre is a set of practices and strategies that develop and are widely recognized within a
group of interlocutors (Orlikowski, Yates, 1994; Yates, Orlikowski, 2002; Yates, Orlikowski,
1992; Yates et al., 1999). Two communicative genres with the response network emerged
from our coding (for more detail on the definition of the response network, please refer to
chapter 3.1). While a fervent communicative genre was commonplace among field actors,
administrative actors primarily interacted in a dispassionate communicative genre. The
“fervent” communicative genre included a great deal of discussion as well as some personal
and emotional involvement. The dispassionate communicative genre developed around emails
and faxes and implied that interlocutors focus on objective data with low personal and
emotional involvement.

  238
Improvisation rapidly grew organizational but its expansion was limited and did not
encompass the whole response network. Improvisation became organizational in hospitals and
some institutions. However, crisis responders that mainly communicated through emails in
the dispassionate communicative genre had a restricted participation into organizational
improvisation. Figure 48 presents the overall view of the proposed framework. We then detail
each part of the framework.

Figure 48. The overall framework

3.I. Organizational emptiness and organizational improvisation

In this study the notion of organizational emptiness refers to the absence of operation of
resources or means in spite of their being necessary to perform a specific process. The notion
of organizational input involved in this definition is two-fold, referring to organizational
resources and means. Barney provides a fair definition and a typology of various resources
that an organization can miss (1991). Following his approach we view two main types of
organizational resources: 1) material such as objects and people or 2) immaterial such as
competencies, cognitive frames and information. We rely on Mintzberg’s definition of a
organizational means (1979). They are of six types: processes, outputs, instructions, informal
connection between individuals and groups, norms and skills. We provide in the coming lines
more details and examples of organizational emptiness.
Following a realist approach, we propose that the concept of organizational emptiness be both
intransitive and transitive (for more detail on the transitivity please refer to chapter 1.3). First
of all, organizational emptiness is an intransitive reality. A mismatch between resource
allocation and the individuals’ needs is an objective fact. For example, the number of deaths
objectively exceeded the capacity of mortuary chambers. Similarly, no information system
devoted to inter organizational coordination existed in summer 2003. However, individuals
also make their own experience of emptiness. Facing emptiness, individuals and groups
experienced a feeling of urgent lack, which relates to the experience of cognitive dissonance,
as defined by Festinger (1957).

  239
Being both intransitive and transitive, organizational emptiness has an independent existence
that we can hardly understand. Even though individuals perceived emptiness through the
experience of crucial lacks during the heat wave; emptiness has an independent existence
from human subjectivity. Individuals’ perception of a lack of resources or means does not
mean that these resources and means do not actually exist. In the aftermath of the heat wave,
crisis responders eventually understood that some resources and means that they missed at the
time of the crisis response actually existed. For example, hospitals experienced shortages of
medicines despite considerable stocks. In this case, emptiness stemmed from the rigidity
threat in relation to resource supply processes.
Improvisation helped to partially fill the organizational emptiness by providing ad hoc
solutions. For instance, firefighters squads and the SAMU personnel developed an ad hoc
coding system to distinguish the deaths that they suspected to be due to the heat from others.
Improvisation as a response to emptiness has already been reported in literature by several
authors. The ambiguity and lack of definition of some processes can provide an opportunity
for actors to develop unexpected behaviors. Burt, who founded the theory of structural holes,
explained how ambiguities in human agency could raise unexpected behaviors such as
opportunistic strategies (Burt, 1992, 2004). This connection is consistent with French
sociology that accounts for emptiness as an opportunity for actors to develop their reflexivity
(Alter, 2000) and power (Crozier, Friedberg, 1977). In particular, the garbage can theory
depicts how frequent ambiguity in processes is and how it causes unexpected interactions
among decision makers. As explained by Cohen, March and Olsen (Cohen et al., 1972):
“Situations of decision making under goal ambiguity are common in complex organizations.
(…) Often, problems are resolved without recourse to explicit bargaining or to an explicit
price system market”. (p. 1).
Finally the connection between emptiness and improvisation was conceptually addressed by
Hatch (1999). According to this author, actors cope with structures ambiguity mainly by
improvising ad hoc solutions and meanings. The following paragraphs present pieces of
evidence of organizational emptiness. For each, we demonstrate how actors improvised to
respond to emptiness.
3.I.1.Material resource
All sorts of material resources were missing during the crisis response. More specifically
medical personnel were in need of beds, ice, air conditioners, etc. Other kinds of material
were needed to cool down rooms because the building structures were not adapted to such
high temperatures. As hospitals could not command air conditioners, nurses put them on
wheels to cool down more than one room at a time. Otherwise, nurses put bed sheets into
water and hung them up at windows in the hottest rooms.
Some of the available resources were not usable in some cases. In many hospitals the water
was too hot to be drunk or too refresh patients. As an ERT describes, medical students
improvised by asking for ice to restaurants and bars that were located nearby the hospital:

  240
“I asked for bottles of water. My superior did not understand why I was urging her so much
to get it. Yes, we had some water. But it was too hot to be drunk. So students who had come
back from their vacation asked in bars for some ice. Some barmaid told me: “It is the only
thing I can do”.
3.I.2.Immaterial resources
Information was one of the most needed intangible resources. At any moment, the
administrative sphere could trace events in hospitals. Although scarce, social ties did exist
between administrative and field actors (more detail in the Communicational connections
between individuals and groups section). An information system existed – the PMSI – but
would not provide information in real time. No system enabled the tracking of patients to
detect the work overload in hospitals at that time. Consequently, the response network
subsequently had to establish an ad hoc information system to report their daily activity. As a
DGS manager described the settling of an ad hoc information system:
“As requested by the [Health] Minister’s advisers, we required from the ARH that all the
hospitals adapt to the situation. We were informed that the situation was exceptional. At that
moment we asked every hospital director to transmit information. We stressed our need to get
information on a daily basis. This is why we provided a form that we transmitted to the ARHs.
Frequent phone interactions were helpful to forecast tendencies on a short notice.”
Organizations were missing cognitive resources to cope with the situation as well. During
public hearings, an officer from the BSPP explained that there was no code to refer to the
specific situation they were facing, which prevented firefighters from following established
guidelines. As this officer explains, firefighters had no idea of how they could handle the
situation. For this reason, they improvised a new code.
“We had to invent a new code, this code was called “syndromes due to heat”. We had not
thought about this code previously”.
3.I.3.Processes
During the French heat wave, squads that were intervening did not know what type of
intervention they were completing. Therefore, they had no idea of what information they
should collect and what parameters were important for decision making. As an officer
explains:
“Teams leaders were asked to analyze each case they handled on their own, without referring
to code protocols. Therefore each case was approached empirically once we understood there
was a significant number of interventions due to fainting”.
To respond to this lack, the BSPP squads developed ad hoc criteria for decision making. More
than developing a new intervention code, firefighters created an ad hoc process to handle data
in relation to hyperthermia:
“Since the 7th of August, officers were asked to complete their reports with complementary
information such as environmental and body temperature and aerologic conditions. By doing

  241
so, they had a better idea of the role heat played in fainting patients and made better
decisions”.
As a firefighter officer explained, the BSPP squads cooperated with the SAMU to settle the
ad hoc process to take care of the victims of hyperthermia:
“Since the 8th, we cooperated with the SAMU to develop therapy procedures to take care of
the victims ourselves and alleviate the workload in hospitals. We created what we called a
refreezing protocol, a reanimation protocol and a medical algorithm to decide whether to
transfer the victims to hospitals or not (…)”.
3.I.4.Instructions and orders
Institutions had to provide new directions so that resource shortages could be resolved almost
everywhere. Resources were missing because directions to get resources had not been
defined.
“There were sudden shortages of resources in many facilities. In collaboration with the DGS
we gave directions so that hospitals would help assisted living facilities manage shortages.
The situation was unusual. We enabled drugstores to get more medicines to organizations
that were in critical situations”.
Similarly, a specific instruction in relation to alerting was lacking. As an operational actor
explained during a public hearing:
“I would have been able to transmit data and alerts. But to who? I have no supervisor to
contact in case of problems”.
During crisis response, ad hoc instructions emerged and were transmitted by fax from
institutional actors such as the AP-HP. These instructions recommended the postponement of
scheduled activities to focus efforts on hydrating patients who were suffering from
hyperthermia.
3.I.5.Outputs
Within the response network, some organizations could not reach their objectives and could
not fulfill their missions. For this reason, these organizations had to redefine their outputs. For
instance the INVS could not complete its health-watching mission due to the lack of
information. As the INVS explained during public hearings:
“The InVS mission is to inform the DGS about all the potential threats, such as meningitis,
legionaire’s disease, SARS”.
Similarly, the medical personnel and civil protection organizations were unable to fulfill their
mission of saving lives when the number of victims suddenly increased. In these cases, the
crisis responders redefined their tasks into bringing ice, hydrating patients, counting victims
and sending faxes.

  242
 3.I.6.Norms
Crisis responders faced unexpected situations in relation to which no rule or norm preexisted.
Because of that, they developed divergent views, depending on their professional values or
their common sense. The implications of the lack of norms can be illustrated by the dilemma
experienced by firefighters when it comes to transmit information to journalists.
During the crisis there was no predefined norm or instruction about information transmission
about death rates. When contacted by journalists with respect to death rates, a BSPP officer
asked for advice to the head physician. The latter was agreeable to transmit information. As
the BSPP officer who was in charge of communication with journalists explained:
“When journalists contact me to get information about the death rates I go to the head
physician of our squad and I ask him if I can tell the journalists that 7 persons died. He
responded yes, because 7 is the minimum approximation of deaths and that we have nothing
to hide. He is ok to transmit the figures to the journalists but he tells me that we need to
validate this decision by the PPP. When I told him that the journalists wanted to interview me
on that topic he immediately contacted the Police authorities (…)”.
The PPP officer contacted by the BSPP officer disagreed on transmitting information. At that
moment a dilemma tore the firefighter’s mind. On one hand, BSPP officers should abide by
the PPP officers’ decision. On the other, the public expected to be informed about abnormal
death rates. This dilemma led him to deviate from established obedience due to superiors in
the military culture: He personally expressed his view and attempted to negotiate with the
PPP officer. The improvised attempt for negotiation aborted. However, the lack of norms led
crisis responders to deviate from established patterns of behavior, which is a core feature of
improvisation (Clegg et al., 2002). As the BSPP officer explained:
“My boss calls them while I am standing into the office. He says that the situation is critical.
He requests permission to transmit information about the deaths that are due to the heat
wave. He then switches the phone loudspeaker on, which allows me to hear the following
instructions: We have to avoid any public psychoses and therefore we should not transmit any
data. (…) I heard my boss having the following conversation with the PPP officer:
- You should know that the whole country gets concerned when one homeless person
freezes to death in the winter. Currently, more than seven persons have died from the
heat. This may be an even more important situation today…
- Yes but we are not really sure that these people died from the heat, are we?
Then my boss nodded. (…) To me, physicians are the most competent to decide whether they
could transmit 7, 8 or 40 as figures”.
3.I.7.Skills
At the beginning of the crisis response, physicians were convinced the symptoms of
hyperthermia were due to some infectious disease. For this reason they gave antibiotics to
their first patients. Physicians were powerless because whatever they attempted to do, patients

  243
would die in their hands. Specific know-hows to manage the effects of hyperthermia were
lacking. As an illustration of this lack, one of these physicians spent an entire Friday night on
the Internet searching for information about hyperthermia and its effects:
“[A physician] collected research on hyperthermia on the Internet. He already knew that heat
can kill but did not understand how. On Friday, he used free access to research documents
and read over everything. Then he discovered the same thing had happened in Chicago in
1995” (Jacquat report, pp.283-289).
Hospitals’ staff similarly responded to the lack of skills and competencies in relation to
hyperthermia. Improvisation, as a trial and error process (Ciborra, 1996b), enabled diverse
actors to develop know-how with respect to hydration.
3.I.8.Communicational connections between individual and groups
Not only were information systems disconnected but individuals and groups would miss
communication tools such as common vocabulary and codes as well. They were also reluctant
to share information or respond to each other’s alerts. This dimension of organizational
emptiness is not new in the crisis management literature (Garnett, Kouzmin, 2007) and the
management literature. In their study on collaboration in virtual teams, Panteli and Fineman
explain how silence is interpreted by actors as a message and plays a full part in
communication in organizations (2005).
It is important to note that such communicational gaps were inherent to the structure of the
Parisian Heat wave Crisis Response Network (HCN). Furthermore, due to the diversity of
organizations and functions within the network, actors experienced the heat wave in radically
diverse ways. Operational actors were struck by the heat, the feeling of being powerless
against death and the lack of water. In the administrative sphere, the feeling of crisis was less
brutal. This same feeling was delayed as actors got information about the situation, urgent
calls, or watched Patrick Pelloux’s intervention on television on the Sunday evening news.
The lack of communicational bridges is illustrated in the role that information retention
played in crisis response. As some INVS director explained, field and administrative actors
exchanged little information:
“Between the 6th and the 8th of August, transmitted information is parceled: only one DDASS
– and I am not accusing anyone, this are only facts – from Morbihan (Nota: A French region
on the west coast) informs us that there has been tree deaths of pretty young people. That is
all”.
Another example of missed communication bridges came from the head of the Health
Ministry advisers. As field actors were experiencing the catastrophe in real time
administrative actors experienced the crisis indirectly and did not always make sense of the
situation as a crisis. In particular, they hardly figured that the situation was critical, partly
because the words “crisis” or “heat wave” were not clearly mentioned. As one of the health
ministry advisers explained:

  244
“Before I arrived on the 11th of August I had received no phone call during my break. I had
300 emails to read and no one clearly mentioned the heat wave”.
Organizational improvisation helped to fill, at least partially, the lack of communication. As
an administrative director explained: 
“Today, men and women who participated in the response to the heat wave can be proud of
having given everything that they could give at that time. This collective struggle developed a
belonging feeling and solidarity among teams. The most frequently used attribute to depict
the experience of the heat wave is « federating »”.
Figure 49 summarizes the mutual influence between organizational emptiness and
organizational improvisation. Arrows represent crisis responders’ improvised practices as a
response to organizational emptiness.

Figure 49. Organizational emptiness and improvisation

  245
 3.II. Communicative genres during the 2003 French heat wave
We detail here the notion of communicative genre and expand on their impact on
organizational improvisation. Some preliminary codes about technology use emerged during
analysis. We eventually refined these codes by relying on existing studies on the concept of
communicative genres. By doing so, we identified two communicative genres that coexisted.
We first define the notion of communicative genre. Then we detail the fervent and
dispassionate communicative genres. The two communicative genres are depicted in Figure
50.

Figure 50. Dispassionate and fervent communicative genres

Dispassionate Fervent communicative

communicative genre genre

Discussing micro
Process narration issues

Data hunting
Opening up to new
ideas

Centered
communication

Embracing others'
experience
Protecting from
hierarchical
disapprobation

Orlikowski and Yates’ work provide a clear explanation of what a communicative genre is
(1994). As those authors explain:
“A genre of organizational communication such as the business letter, shareholders’
meetings, or report, is a distinctive type of communication, characterized by a socially
recognized and common aspect of form (…). The communicative purpose of a genre is not
rooted (…) in a purpose that is collectively constructed, recognized and reinforced within a
community” (p.543).
In other words, a communicative genre consists of a repertoire of communicative practices
and strategies that are associated with specific social codes and communication media. For
instance business letters correspond to specific communication strategies towards
shareholders. Similarly to routines, genres develop on the long term.

  246
 3.II.1. The fervent communicative genre
The fervent communicative genre describes vivid exchanges between responders, including
emotions and controversies. When they use the fervent genre, the crisis responders generally
interact in face to face or by phones. In addition, they aim to persuade and emotionally move
their interlocutors. We detail the fervent communicative genre features and how the fervent
communicative genre supported the development of organizational improvisation at a
collective level.
3.II.1.1. Discussing micro issues
When communicating in a fervent genre, crisis responders discussed specific events or issues.
Focusing on micro issues, interlocutors were more prone to express their personal opinions
and discuss them. For example, a manager from the DHOS explained how she expressed her
concern to the government about a specific AP-HP alarming message sent of the 8th. By doing
so she explained the Health Ministry’s advisers that the meeting that was to take place on the
11th was an opportunity to respond this message:
“The central administration does not usually participate in the hospitals’ internal meetings.
But this time I warned the ministry’s advisers that the message I had received from the AP-
HP was too alarming. Then one of the advisers told me that I should attend the meeting that
we were invited to. She told me that I should ask the AP to invite the DGS”.
This kind of interaction permitted dialogic reasoning between actors. Dialogic reasoning is a
process during which actors develop their ideas collectively and dialectically (Faraj, Xiao,
2006). Dialogical reasoning is reported to happen often in organizations that need to cope
with critical situations. During the 2003 French heat wave, dialogical reasoning frequently
occurred to find ad hoc solutions.
Actors’ involvement in crisis response depended on their ability to discuss to develop a
common understanding about the situation. For example, cases of old patients who assail
physicians with requests are typical in routine situation. For this reason, physicians
underestimated the suffering of some old patients in their services at the beginning of the heat
wave. Hospitals’ medical personnel overcame misrepresentations about the elderly mainly by
discussion:
“On the 11th, I understood one thing: to mobilize all services and not only emergency services
in hospitals we had to explain that the situation was distinct from the ordinary issue of
abandoned elderly. The old person with the SAMU phone number anchored next to the phone
is a cliché. At that time field actors who went into emergency rooms saw what was happening.
This is why I argue that we need to see and talk”.
Moreover discussion was definitely needed to be reactive. As some administrative actor
points out:
“We would have needed to confront several sources of information. We needed data from the
field; all the data that comes from hospitals and restrooms because these were the places

  247
were the number of deaths was its climax. We should have brought information from these
services as well as others to a single point”.
Through discussion, responders developed dialogical reasoning with experts, supporting the
emergence of expertise leadership. Both dialogical reasoning and expertise leadership can be
viewed as a core component of improvisation. As a BSPP officer explained, discussion with
physicians within the squads provided the opportunity for officers to understand from the
physicians that the heat wave was likely to have an impact on people’s health. Furthermore,
the discussion led the physician to lead improvisation with respect to the measurement of
temperatures among firefighters:
“By discussing with my superiors and physicians, I noticed that they were trying to make a
connection between the growing number of victims and the heat. The physician took the first
measures of temperature and then asked everyone to do the same with every victim”.
3.II.1.2. Opening up to new ideas
The fervent communicative genre is characterized by the interlocutor’s feeling free to tackle
unexpected topics. Interlocutors started some subjects. As conversations were occurring,
various topics, problems or solutions were brought up. This kind of interaction is
commonplace in organizations and was described by the Garbage Can theory (Cohen et al.,
1972). Interactions gather issues, solutions and ideas in a pell-mell fashion. Creativity is more
important during such interactions because interlocutors make unexpected connections
between problems and solutions.
Sometimes new problems were mentioned during the conversation, leading actors to
brainstorm together and develop a new idea. During public hearings, the AP-HP director
explained how she had the idea of calling students back:
“On the 11th of August, I was discussing with a service care director. We were wondering
what we could do to cope with the situation. Then we thought we could call students to
volunteer (…). As a result 400 students came back (…). There was so much to be done”.
In other cases, the interaction was provoked to solve a specific and concrete problem. Patrick
Pelloux’s essay provides a fair description of brainstorming meetings that would open to new
ideas as sources of improvisation and bricolage. For instance, he depicts how administrative
actors had the idea of using food vans to stock bodies so as to alleviate mortuary chambers
that were overloaded.
Finally, some of these interactions did not necessarily provide new ideas but enabled actors to
be more vigilant about some potential issues and transmit their concerns among colleagues. In
this kind of situation, actors were more prone to improvise because they had in mind the
absence of a predefined solution. As an ERT explained:
“This crisis meeting gathered specialists from various disciplines among which the AP-HP
ones. Deviating from the original subject, summer potential issues such as bed were
mentioned aside”.

  248
Opening up to new ideas created room for spontaneous suggestions and innovating ideas,
therefore supporting improvisation.
3.II.1.3. Embracing other’s personal experience
Actors built a common vocabulary and shared their experience in an empathic manner. This is
consistent with previous findings on the importance of boundary spanning in crisis response
(Kapucu, 2006). Actors took into consideration personal emotions and concerns from
colleagues and counterparts. As some administrative actors described her reaction after her
discussion with a counterpart:
“ At 8 o’clock in the morning, I receive this call from AP-HP. I am driving at that time. But I
stop my car at 9 o’clock because at Porte d’Italie to call back because I had found this call
disturbing. I called to say that I needed to attend the meeting because the alert from [my
counterpart] – that I knew very well – seemed very alarming”.
Embracing other’s experience participated into discussion and dialogical reasoning. As a
BSPP officer explained, sharing about one’s feelings contributed to build an ad-hoc alert
threshold:
“I think the alert threshold was surpassed because my services twice alerted upper services.
This alert threshold is not something mathematical. It is rather a general feeling that stems
from so many interventions that a particular event”.
Embracing other’s experience implied the importance of inter individual connection and
knowledge sharing. It required that interlocutors provided details about their own experience:
their sensation, emotions or even the people or objects they were surrounded by. For instance
there was the need in some hospitals to get some fresh air. Some ERT described the
interactions that took place between all the people who tried to cool down rooms for patients:
“Plumbers, firefighters, building workers and so on. They all deployed the creativity and
spent hours to think about possible bricolage to cool down the place. They tried to improve
air channels, they sprayed water on the building walls, they covered the walls with bed
sheets”. (l’Urgentiste, p.55)
Furthermore the fervent communicative genre implies that individuals felt confident enough
to express their emotions:
“The person who received alert calls on the 7th of August knew very well how hospitals work.
His knowledge enabled him to understand his interlocutors’ situation and made him think that
something very serious was coming out. This is why he immediately called [his colleagues]”.
This is why empathy and trust were core aspects of the fervent communicative genre. Such
interactions implied a frank expression, with no euphemism. An ERT reported his discussion
with a colleague. At the end of this conversation, the two colleagues deviated from
established patterns of actions by directly calling national organizations directors (DHOS and
INVS) to alert them:

  249
“This night, information from colleagues converges: people are dying in hospitals halls. (…) I
then receive this call from this professor:
- I was about to call you, I told him. Have you heard about all that stuff? Things are
bad…
- I think we are in deep shit, he answers.
At that time, we mutually agree to act. We decide to alert. He decides to take care of the DGS.
I will handle the DHOS”. (L’urgentiste, p.38).
Embracing other’s experience promoted spontaneity and boundary spanning processes that
are core features of improvisation.
3.II.2. The dispassionate communicative genre
In this section we describe the dispassionate communicative genre and its effects on
organizational improvisation. As a reminder, the “dispassionate” communicative genre was
developed in relation to email and fax use. Two examples of email archives can be found in
the Annex, in Figure 54 and Figure 55.
3.II.2.1. Process narration
Actors primarily reported ongoing processes and data in emails. Processes were mentioned to
introduce what tasks were needed. For instance, on the 6th of August, one of the Health
ministry advisers explained to his counterparts at the DGS his concern about the effect of the
heat on citizens’ health. He then suggested the need for a recommendation memo about the
heat wave that should be sent to health care organizations:
“It would be useful to send a memo to recall basic precautions to care for the youngest and
eldest patients. There are plenty of studies about the health related impacts of heat waves
(…). I think the CDC [Center for Disease Control in Atlanta] reviewed them). It could result
into an emergency message”.
Actors spent three days responding to this initiative. As Patrick Pelloux explains in his essay,
they exchanged 60 emails to end up with a final version of the memo. Interaction via email
reported major or minor revisions from interlocutors on the document.
Otherwise emails were used to transmit information about what had been completed
previously. By doing so actors developed a narrative structure that supported coordination.
For instance, an INVS 22employee reported the general ideas that emerged from an informal
phone interaction with colleagues. He sent his colleagues the following email:
“I called [colleagues] to discuss about what could be done [to investigate the health related
impacts of the heat wave].

22
INVS is the national health-watching institute in France. For more detail on the response network
organizations please refer to chapter 3.1
  250
First of all, we should take into account the feasibility criterion. We therefore should
collaborate with organizations that already have an information system and data from
previous years
Here is our idea: trying to bring information up to the INVS rather than settling for a specific
data collection system”.
By mentioning a previously received email, the sender transcribed what was discussed
previously to his email, which supported process narration. The email was then forwarded
multiple times without any comment or additional idea. Its content was of the strategic level,
dealing with feasibility as a generic criterion to make decision on whether the strategy should
be implemented. However, the interlocutors did not mention any detail on how to implement
the strategy. It is interesting to note that the INVS never implemented its plans for
investigation.
Despite the development of a narration, discussion to implement ideas scarcely occurred
when crisis responders communicated in a dispassionate genre. Process reporting lessened
opportunities for dialogic reasoning between email and fax users. These users hardly
approached micro issues as well, even though phone exchanges afforded the opportunity to
talk about ongoing actions.
3.II.2.2. Data hunting
When objective, data provides a legitimate basis to make decisions. For this reason, email
users extensively sent figures and data during crisis response. As a result, administrative
actors spent the whole crisis response period to find out the number of deaths.
Indeed emails including data were systematically transmitted between actors, but with scarce
discussion. Moreover they hardly mentioned concrete facts, objects and persons. For instance
the following message was initially sent to the INVS and then transferred to seven actors:
“There was a call at 4 pm from the DDASS in Morbihan. Three persons had died at work. As
reported vby the emergency services, these persons were pretty young. The person called the
DGS and was advised to call the INVS. Here is his phone number (…)”.
First the message was introduced in a forwarded email as follows: “I let you evaluate the
situation”, or “FYI”.
Then some interlocutor announced common points and differences between dead persons
(such as age, weight, gender) to evaluate whether these deaths were due to infection or not.
The objective of these emails was to negate or confirm the interpretation of the situation by
the interlocutors with more data. Actors strove to stay as neutral as possible when it came to
interpretation and to refer to objective figures. They rarely expressed personal views of the
situation or refer to their sensations. In many cases, interlocutors expected each other to make
sense of the data, which delayed action and obstacles extemporaneous action (Tylern, Tang,
2003).

  251
As a core feature of the dispassionate communicative genre, data hunting prevented actors
from fueling reflection with opinions and emotions. As data is central in interaction, there was
little room for discussion.
3.II.2.3. Centered communication
Remaining focused on the reporting of administrative processes, administrative actors hardly
integrated field actors’ requests for help or action into their process. For instance an ERT
contacted one of the DGS employees on the 11th of August. His email vividly depicts his
feeling powerless to tackle deaths and suffering:
“Many elderly die in emergency rooms indirectly or directly from hyperthermia. Three
persons died in Saint Joseph [hospital] so far. The emergency roomers (ERs), at least ours,
are totally flooded with patients. Some of them stay 4 or 5 on the same stretchers in ER. We
have to search for more beds and have completely deleted all our scheduled medical
interventions. I have never experienced such a situation in 25 years. The situation is really
serious”.
His administrative interlocutor reformulated his request as a transmission of information and
did not respond to the ERT’s opinion on the criticality of the situation. His response did not
include any personal reaction to the description of the suffering and his tone remained neutral.
He accounted for forms as management tools to be used in hospitals in generic terms:
“Thank you for the information you transmitted earlier about heat strokes in Saint Joseph and
Bichat hospitals. I relayed it internally. The InVS has just settled an emergency data
collection process. Please find attached the forms that compose it. All these documents have
been sent to hospitals. The person to contact is….”.
The ERT response unveiled statistics issues related to data collection:
“Thanks for your message. We have sent messages everywhere. The problem with data is that
they will not reflect indirect effects of hyperthermia (…). The situation is desperate here. The
government or the DRASS should trigger the White Plan or an Emergency plan very quickly.
As I made my promise to your colleague, could you please transfer information?”
With no surprise the administrative interlocutor effectively transmitted the email but no
discussion was developed on data collection. It is interesting to note the email response to the
ERT who had alerted administrative actors to get resources:
“Thank you for information. Beyond the current investigation we intend to lead a long term
investigation on the total mortality – or even morbidity – that will take into account some
specific pathologies, such as cardiac, vascular, and respiratory diseases. Our project has to
be validated but we hope that an international comparison will be feasible”.
Hunting data led email users to hold on to a functional interpretation of the situation. It led
them to elude some practical and logical issues. Furthermore centered communication
prevented actors from discussing micro issues and from developing pell-mell interactions.

  252
 3.II.2.4. Protecting from hierarchical disapprobation
In relation to the dispassionate communicative genre, email media was an instrument to show
transparency off. Basing their rationale on risk taking and the traceability of emails, actors
used emails as a traceable proof of their responsiveness. However, the intention to enhance
one’s reactivity resulted in extensive forwarding but scarce discussion. To protect from
disapprobation, email users forwarded their emails to their colleagues, counterparts and those
who were likely to evaluate them. This finding is consistent with recent studies on the
importance of status in electronic communication (Dabbish et al., 2005) and the side effects
of electronic communication (Courbon, Tajan, 1997).
In our view, email use illustrated the actors’ intent to prevent criticism. Therefore they
avoided to mention controversial content, which limited discursive exchanges about the
widespread bricolage and improvisation practices. Similarly operational actors have been
reluctant to use email since they improvised. However tracking interactions is technically
permitted by email storage on server information traceability (Ramesh, 1998; Ramesh et al.,
2002). As highlighted by Car and Sheikh (2004):
“The written record of email consultations enables close monitoring and evaluation of
appropriateness and safety. Whereas face to face and telephone consultations are rarely
recorded verbatim (typically being documented with only a few key words), email provides
direct evidence of patient-doctor conversation” (p.441).
Traceability is also enacted by email users (Boukef, 2005; Ducheneaut, Bellotti, 2001; Kalika
et al., 2007). As a conclusion, we posit that search for protection from potential hierarchical
disapprobation led email users to avoid mentioning controversial aspects of the situations or
specific issues. In addition they remained on a general stance and directed their messages
towards status rather than persons. Figure 51 presents the effects of the fervent and the
dispassionate communicative genre on organizational improvisation. In this figure, the arrows
represent causal consequence relationships.

  253
 Figure 51. The fervent and dispassionate communicative genres and organizational
improvisation

  254
 3.III. Moving from the dispassionate communicative genre
Some administrative actors participated in organizational improvisation during crisis response
in diverse manners. Simultaneously these actors developed new communication practices and
strategies in relation to fax use. More specifically, they personnally involved in messages by
expressing their emotions and opinions. In addition, they expressed their empathy towards
improvisers.
For instance, one of the AP-HP directors spontaneously sent a fax on the 8th of August to
transmit guidelines to hospitals. These directions mentioned the need to alleviate the
workload of emergency units, by sending back homes the patients whose health state was
satisfactorily and canceling previously planned surgery operations. Given the similarities
between these guidelines and the ones defined in the emergency plan called the “White Plan”,
the AP-HP directors unofficially enacted the White Plan in hospitals on the 8th. By enacting
an emergency plan without asking permission to the prefectures or to the Health Ministry, the
AP-HP directors deviated from the rules that health care and civil protection organizations
usually abide by. The AP-HP co-director, who was the author of this fax, wrote a personal
comment next to the guidelines. In this comment, he expressed his concern and his being
aware that hospitals are facing problems:
“The situation is very alarming. I insistently ask everyone to take all the necessary initiatives
as soon as you get this message. I am aware that the situation is complicated”.
This fax message differed from the ones previously sent in that it does not correspond to the
dispassionate communicative genre. Thi message is presented in Figure 52 in the Annex.
Similarly, the DHOS director sent a fax on the 12th of August to recommend that hospitals
feel free to “consider every opportunity to increase beds capacities, even by collaborating
with funeral homes”. In other words, the DHOS director gave permission to hospital directors
to improvise ad hoc solutions to manage corpses.
The DHOS director spontaneously joined the following request to the hospitals’ director in
his message, radically diverging from the dispassionate communicative genre:
“The medical teams as well as all the hospitals personnel have been assailed with requests.
We are currently experiencing an extraordinary situation. In such conditions, the hospitals’
personnel have been demonstrating the best of our values: professionalism, availability,
solidarity. They are the pride of our hospitals and of our health care system. I ask you to
transmit my congratulations and my thanks”. This message is presented in Figure 53 in the
Annex.
Our analysis of these two messages suggests that some administrative actors managed to
participate into organizational improvisation. Thus, they used faxes to communicate and
moved from the usual dispassionate communicative genre to interact with improvisers.

  255
 Section 4. Discussion of findings

This paper addresses the need to understand what makes technology a potential obstacle to
organizational improvisation. There is no doubt that ICT significantly helps resilience and
reactivity in crisis response. However, crisis management authors provide enough evidence to
question ICT support to crisis improvisation. In addition, the theories of fit require further
investigating before applying to cases of crisis improvisation. Following the grounded theory
method, we propose an integrative explanation of the influence of technology on crisis
improvisation. As improvisation constitutes a response to organizational emptiness, there is
the need to promote a “fervent” communicative genre. These are vivid exchanges that involve
personal experience, emotions and objects. They are characterized by discussion on micro
issues and dialogical reasoning. Our analysis unveils that the “dispassionate” communicative
genres developed by the administrative actors of the crisis response network hindered such
interactions. The way users appropriated emails and faxes prevented them from creating a
space for the expression of emotions, disorders and creativity. However, some administrative
actors managed to participate into improvisation. When they did, they gave permission and
guidelines to improvise through faxes. To do so, they adapted their communication practices,
moving from the dispassionate communicative genre to interact with improvisers. Table 44
compares ICT use, communicative genres and participation in organizational improvisation.

Table 44. ICT tools, communicative genre, participation in organizational improvisation

ICT tools Communicative genre Participation in

organizational
improvisation
Emails Strictly dispassionate None
Faxes Dispassionate. However, fax They gave permission and
senders embrace guidelines to improvise
improvisers’ experience and
involve their opinions rather
than staying neutral
Phones, meetings Fervent Full
Our study reveals that ICT tools such as emails, phones and faxes had different influences on
organizational improvisation during the 2003 French heat wave crisis response. We explain
these differences by the fact that crisis responders developed diverse communicative genres
that existed in relation to these tools.
However, the validity of our findings should be discussed. According to the grounded theory,
evaluation is critical process (Wells, 1995). Corbin and Strauss enounce five criteria for
evaluation of grounded theory studies (2008, p305-30) that refer to: i) the quality and the
reliability of data, ii) the consistency of our research as a process, iii) the applicability of the
research outputs.

  256
First of all, we triangulated the three sources of data to grant a minimum of data reliability. To
validate the consistency of our research process, we focused on the quality of presentation of
concepts, logic of reasoning and relevant contextualization (Corbin, Strauss, 2008, p. 306).
Iterative discussion with colleague, senior researchers and practitioners who were specialized
in IS, organizational theories and health care management enabled us to meet these criteria.
By applicability, Corbin and Strauss mean the usefulness of the findings and the extent to
which they offer useful insights to participants and researchers (Strauss, Corbin, 2008). Our
findings suggest some managerial recommendations in relation to email use that we discussed
with practitioners, which supports a minimal level of applicability of the research outputs.
Finally, we assessed the reliability of our findings, as suggested by Gombault (2005, p. 51).
We relied on triangulation and abduction to test the coherence of our theoretical framework
(De Vaus, 2001). Finally, attaining theoretical and semantic saturation was a technique to
grant the completeness of our findings.

Conclusion

Our study provides some additional insights on the conceptual ties between organizational
improvisation and ICT use in crisis response. Crisis responders’ participation in
organizational improvisation can depend on their communicative genres. Based on empirical
evidence, this study suggests the importance of interactions and communicative genres when
crisis responders have to improvise. Our findings suggest that because of their ICT use, some
departments and groups within an organization can stay away for improvisation in spite of
having great the potential ability to facilitate improvisation.
This study addresses the relevance of applying the theories of fit to assess ICT support to
crisis improvisation. By doing so, it calls for further investigation to enrich the theories of fit
in crisis response.
Our findings are consistent with recent academic concern about organizations’ reflexivity in
relation to ICT use. The 2003 French heat wave illustrates the risk of transforming our means
of communication as cold and formal channels, where unexpected news or data remain out of
the frame. Likewise some authors stress that electronic communication mismanagement is
likely to burden organization’s processes (Isaac et al., 2007; Kalika et al., 2007; Tran, 2010).
This paper brings more insight on the organizational resources that are likely to influence
improvisation. Other authors have stressed the importance of organizational resources in
improvisation, such as organizational memory and information flow (Moorman, Miner,
1998a). However, additional investigation would allow clarifying the relationship between
organizational memory and communicative genres.
To conclude, it is important to note that this study does not question the appropriateness of
email in crisis response. Examples of efficient use of emails in crisis response are abundant in
the literature (Jefferson, 2006a, b; Marincioni, 2007). Rather, we highlight the need to
manage our means of communication to enable interactions when a crisis occurs. In

  257
particular, we suggest the need to create virtual spaces that can play the role of garbage can
where users feel free to provide new ideas. We expect this research to help managers to
develop a better use of technology during improvisation. Interestingly enough, some
improvisers managed to adapt their communicative genres around email and faxes. For
instance, some of them wrote by hand messages on faxes. By doing so, they mentioned their
own action, emotion or opinions and departed from the dispassionate communicative genre.
By doing so, they made the content of the message suitable to crisis improvisation. For this
reason, we practically recommend to managers to promote other communicative genres than
the dispassionate genre in relation to a single media. We further detail managerial
recommendations in chapter 4.2.

  258
Annex

Figure 52. An AP-HP fax message

  259
 Figure 53. An DHOS fax message

  260
 Figure 54. An example of email internal
archive

  261
 Figure 55. Another example of email archives

  262
  263
 Part 4. Conclusion part: Summary of findings & final
discussion
The conclusion part summarizes the findings of each study, puts into perspective these
findings and looks ahead for new research challenges. This part is composed of two chapters.
In Chapter 4.1, we present a summary of our findings and then detail the final findings of the
dissertation.
In chapter 4.2, we discuss the contribution of our work. We first suggest some managerial
recommendations in relation to ICT use for crisis improvisation. Then we present the
theoretical implications of our work. Finally, we examine some limitations of our work and
introduce new avenues for research.

  264
Chapter 4.1. Dissertation summary

Introduction

In this chapter we propose a short summary of each of the three articles that compose the
dissertation project. This chapter is structured into three sections: In Section 1, we make a
quick recall of the dissertation structure and its research questions. Section 2 presents a
summary of the findings of each article. Section 3 is structured into two parts. First, we
discuss the connections between each article findings. On the basis of this discussion, we
present the overall findings of the dissertation. We then provide a synthesis of our research
questions and corresponding findings at the end of this chapter.

Section 1. The dissertation structure and its research questions

The primary research question of the dissertation can be formulated as follows: “What
support does Information and Communication Technology (ICT) offer organizational
improvisation during crisis response?”
The management literature provides both pros and cons about technology support to crisis
response. In addition, authors have paid scant attention to improvisation in crisis response, in
spite of a developing body of knowledge (Mendonça et al., 2010). As a result we know little
about how crisis responders interact and coordinate during crisis improvisation. Answers to
the question of whether technology supports crisis improvisation are even scarcer. Finally,
managers lack guidelines to use ICT efficiently and to manage crisis improvisation.
Because technology use, improvisation and crisis are multidimensional concepts, the primary
research question requires further specification. We clarify our understanding of the three key
terms of the primary research question “ICT”, “improvisation” and “crisis” in the introductory
and theoretical parts of the dissertation (respectively Part 1 and Part 2). By doing so, we also
segment our dissertation project into four questions that we address in three studies.
Even though a significant amount of definitions and conceptualizations of crisis are
documented in the management literature, our work is based upon specific assumptions in
relation to the concept of crisis. We approach crisis through the concept of organizational
crisis. In addition we restrict our investigation of organizational crisis to the stages of crisis
response and crisis coping. Finally, we focus our investigation on organizational crises
provoked by disasters.
Improvisation is a polymorphous concept and has been studied in various settings. A large
panel of cases of improvisation can be found in the management literature ranging from
routine strategic decision making in theatre (Ford, 2008) to users’ adaptations to newly
implemented technologies (Brigham, Introna, 2006; Orlikowski, 1996; Orlikowski, Hofman,
1997). Furthermore, authors investigate improvisation from multiple standpoints: as an
organizational paradox (Clegg et al., 2002; Vera, Crossan, 2007), as an innovative response to
unexpected events (Webb, 2004; Webb, F-R., 2006), or even as a collective capability (Vera,

  265
Crossan, 2005). Researchers have experienced increasing difficulties in making sense of the
diversity of academic work on improvisation.
From the analysis of the key terms of the primary research question, we deduce the four
following research questions:
I. How to cope with diversity in research on improvisation?
II. What possibilities does ICT offer for organizational improvisation in crisis response?
III. Does ICT fit users’ needs with respect to organizational improvisation in crisis
response?
IV. How do users benefit (or not) from ICT when they improvise collectively in crisis
response?
We address research question I in the first study. We performed factorial and cluster analyses
on a sample of 105 articles to identify divergences among authors’ approaches of
improvisation.
We address research question II in another study. We analyzed a sample of 15 papers that
deal with ICT use for improvisation in crisis response. To do so, we preliminary deduced
from existing literature some theoretical propositions on the constituents of organizational
improvisation in crisis response.
In a third study, we address both research questions III and IV in a qualitative and
retrospective case study, our last paper project. In this study, we completed a grounded theory
analysis of email, Internet, fax and phone use by crisis responders of the 2003 French heat
wave.

Section 2. Presentation of findings

In this section we summarize each of our findings that resulted from the three studies.
2.I. Coping with diversity in research on improvisation
The first study focuses on the research question entitled: How to cope with diversity in
research on improvisation? In this work we propose that diversity stems from divergences
between authors with respect to four tasks that compose the research process: i) Evidencing,
ii) Focalizing, iii) Contextualizing, iv) Creating knowledge on improvisation. Within each
task, the authors’ divergences stem from the coexistence of two primary perspectives on
improvisation. We list and detail these four tasks and eight perspectives in the following
paragraphs.
First of all, authors employ distinct strategies to evidence improvisation. Some studies
conceptualize improvisation by deductions from previously published definitions and
findings. Conversely, other studies propose an original conceptualization of improvisation
that they induce from empirical findings or abductively discuss.

  266
Secondly, authors focalize on various aspects of improvisation. Some studies deal with
collective improvisation in top management teams that design activities. Contrasting with this
approach, other studies examine improvisation by individuals from multiple hierarchical
layers.
Thirdly, we identify two perspectives in relation to the contextualization of improvisation.
The former presents improvisation as a response to crisis while the latter suggests
improvisation is a way of organizing that supports the development of new capabilities in
routine or turbulent settings.
Finally, authors target divergent knowledge creation objectives. Some studies build on a
unified body of knowledge on the very concept of improvisation. Some other studies aim to
offer alternative views of improvisation, either by challenging existing conceptualizations of
improvisation or by relating improvisation to other concepts in management.

Figure 56. Eight primary perspectives on improvisation

2.II.ICT support to organizational improvisation in crisis response: a literature

review and theoretical propositions
In this study, we aim to respond to the research question II, entitled: What possibilities does
ICT offer for organizational improvisation in crisis response?

  267
In the first part of this study, we rely on the dynamic capabilities approach to make some
theoretical propositions on organizational improvisation. Thus we depict organizational
improvisation as a collective process that involves spontaneous deviation from the established
use of resources and requires specific interactions among actors who command these
resources. Boundary spanning, expertise leadership and minimal structures regulate
interactions that take place between crisis improvisers.
In the remainder of this study, we draw on our theoretical propositions on organizational
improvisation to evidence the influence of some ICT properties on organizational
improvisation in crisis response. First of all, the graphical representation of data during crisis
response creates a unique reference that places shared information into a uniform format
(Comfort, 1993), making of technology a boundary object. In addition, graphical
representation enables fast interpretation and fosters, in our view, spontaneous generation of
ideas or spontaneous adjustment. Secondly, modularity supports the deviation of established
use of resources by enabling the fast implementation of ad hoc interface. Thirdly, the
combination of centralization of data and calculation is a starting point for expertise
leadership. Fourthly, combined virtuality and many-to-many communication supports both
expertise leadership and deviations from established uses of resources. Fifthly, structured
centralization of data enables actors to find information concerning others’ past actions,
therefore promoting boundary spanning interactions and the development of a common
referential. Finally, boundary spanning practices between distant groups cannot exist without
immediate and free information replication and exchange (Comfort, 1993). Virtuality also
helps to keep shared information consistent in time, thanks to information feedback.
Therefore, minimal structures can be initiated and managed by experts (Yuan, Detlor, 2005).
In Figure 57, the arrows represent the positive effect of ICT properties on the constituents of
organizational improvisation in crisis response.

  268
 Figure 57. ICT properties and crisis organizational improvisation constituents

2.III. ICT, communicative genres, organizational improvisation and emptiness

during the 2003 French heat wave crisis response
The third study responds to the limitations of our theoretical work by taking into account ICT
user’s contingency.
Our first objective in the empirical part of the dissertation is to respond to research question
III entitled: Does ICT fit users’ needs with respect to organizational improvisation during
crisis response? To respond to this research question, we assess whether the ICT services that
support interaction between improvisers are adequate. Our point is that the existing theoretical
background on ICT requires conceptual enrichment to provide appropriate responses to
research question III. The reasons that account for the theories of fit not being adequate to
understand whether technology fits to crisis responders’ needs are threefold. First of all, the
theories of fit do not provide enough description of the task characteristics to distinguish
crisis improvisation from other type of tasks. For instance, crisis response in the context of
natural disasters involves numerous individuals from diverse organizations. Similarly, the
theories of fit overlook some aspects of crisis response that are likely to affect ICT use, such
as emotional pressure. Secondly, improvisation is, in it’s essence, unpredictable. As a result,
researchers cannot evaluate the tasks that will be improvised by advance. Therefore, as a
  269
deterministic theoretical background, the theories of fit cannot be directly applied to the
specific case of improvisation. Finally, the theories of fit make a distinction between the
technology characteristics construct and the task construct. However, crisis improvisers can
improvise on ICT. For instance, crisis responders can adapt to the technological tools they
have at hand, even though these tools are not adequate. Due to the lack of “fit” of the theories
of fit to the specific case of organizational improvisation during crisis response, we have led
an in-depth investigation of the interactions of crisis improvisers, thereby addressing research
question IV, entitled: How do users benefit (or not) from ICT when they improvise collectively
in crisis response?
In this part of the empirical study, we investigate the ICT tools that were used during the 2003
French heat wave crisis for interaction, e.g. email, phone, Internet and faxes. We include in
our inquiry mobile phones because mobile devices are commonly used for coordination in
medical settings (Ammenwerth et al., 2000). During the 2003 French heat wave crisis,
organizational improvisation took place but did not develop to the highest scale, e.g. the
whole response network. While some field actors improvised together by using phones, most
administrative actors failed to participate in improvisation despite extensive interaction
through emails. Seemingly, improvisers did not benefit from ICT. They either had a low use
of emails when improvising collectively, or they hardly improvised while using emails and
faxes. Following the grounded theory method, we propose an integrative explanation of the
influence of technology on crisis improvisation during the 2003 French heat wave. Our
analysis reveals that improvisation constitutes a response to organizational emptiness that can
burden crisis response. However, crisis responders’ participation in organizational
improvisation depended on their communicative genres. The fervent communicative genre is
characterized by high emotionality, vivid discussion on micro issues and dialogical reasoning.
This genre supports organizational improvisation. On the contrary, the administrative actors
developed another communicative genre around electronic communication that hindered
organizational improvisation within the network. This communicative genre, that we label
dispassionate, prevented crisis responders from creating a space for the expression of
emotions, disorders and creativity. Table 45 presents a synthesis of the findings from our
three studies.

  270
 
Study title Needs & objectives Findings

1. Coping with diversity in 1) Providing a classification theme to Diversity stems from divergences between authors
research on improvisation compare existing studies on with respect to four tasks within the research process:
improvisation i) Methods to evidence improvisation
2) A clarified view of this existing ii) Aspects of improvisation under observation
perspectives on improvisation
iii) Improvisation contextualization
iv) Knowledge creation on improvisation

2. ICT, crisis and 1) Proposing an explanation of how 1) 5 organizational mechanisms compose

organizational collective improvisation works organizational improvisation in crisis response. Among
improvisation: Some 2) Investigating the role that ICTs play these organizational mechanism, 3 mechanisms
theoretical propositions and in improvisation during crisis regulate interactions within organizational
a literature review response. improvisation.
2) 6 ICT properties promote organizational
improvisation by supporting each of the 6 constituents
of organizational improvisation in crisis response.

3. ICT, communicative 1) Understanding why the 2003 1) Directly applying existing theories of fit does not
genres, organizational French heat wave crisis responders did provide a clear view of ICT fit to crisis improvisers’
improvisation and not use ICT for interactions when they needs.
Table 45. Article projects: needs and findings

emptiness during the 2003
French heat wave crisis
response
improvised 2) The dispassionate communicative genre impeded
2) Determining whether technological organizational improvisation during the 2003 French
tools under use fitted improvisers’ heat wave
needs during the 2003 French heat 3) On the contrary, the fervent communicative genre
wave (FHW) supported the development of organizational
3) Understanding the influence of ICT improvisation
use on coordination in crisis

271
improvisation during the 2003 FHW
 Section 3. Connections between findings

In this section, we detail the logical connections between the three studies. We conclude the
chapter by presenting the overall model that emerged from our work.
3.I. Our approach to improvisation
The dissertation project proposes a comprehensive investigation on ICT support to crisis
improvisation. To do so, we first attempt to make sense of the diversity in research on
improvisation. We rely on our analysis of the management literature on improvisation to
detail our approach of improvisation in the dissertation project, represented in Figure 58.

Measurement and
Reflexive
deduction

Evidencing

Improvisation as a
Individuals' collective process Top management
improvisation to that involves mixed improvisation as a
cope with resources hierarchical status collective design

Observing

Improvisation as a way Improvisation as a

of organizing response to crisis

Contextualizing

Offering alternative
Knowledge building
views

Creating knowledge

Figure 58. Our approach to improvisation

Our evidencing of improvisation follows the reflexive approach. In chapter 2.3 we discuss
existing conceptualizations and their limitations to identify the need to identify the role played
by interactions in organizational improvisation during crisis response. We then make
theoretical propositions that we deductively draw from existing literature. We then complete
an in-depth qualitative analysis of improvisation.

  272
With respect to focalizing, we diverge from the two main perspectives on improvisation by
focusing on improvisation as a collective phenomenon between mixed hierarchical layers.
We contextualize improvisation as a response to critical events.
With respect to knowledge creation, we primarily attempt to offer additional insights on
interactions that compose organizational improvisation and ICT use. Our objective is not to
provide additional insights to the existing theoretical body by offering an alternative view of
organizational improvisation.
3.II. ICT: Three complementary levels of analysis
In the second study, we examine ICT properties at large and examine ICT generic properties.
From this perspective we primarily approach ICT as an set of functionalities that are socially
recognized (Orlikowski, Iacono, 2001)
The empirical part of the dissertation brings more focus on the role played by technology in
organizations. To do so, we evaluate the technological fit to crisis improvisation. We restrict
this part of our investigation to the “interacting” technological function that enables
interactions to respond to the specific needs of crisis responders. We then analyze the
influence of ICTs on crisis improvisation from a practice-based perspective. To do so, we
examine both offline and online interactions of crisis responders. In this part of the
dissertation, we focus on email, faxes and phone use.
Each of these perspectives offers a complementary view of technology. The first perspective
provides some insights on the common effects of technology on improvisation. Our approach
of ICTs is general because our findings result from the use of a common analysis grid to
various descriptions of specific cases of crisis responses. However, this study calls for further
understanding individual and organizational variables that are likely to influence
technological support to organizational improvisation, which calls for the two other
perspectives that we adopt in the empirical part of the dissertation.
By investigating ICT and crisis from two different levels of analysis, we offer empirical
evidence that the technological tools can have distinct effects on crisis improvisation.
Obviously, emails, phones and faxes had distinct effects on organizational improvisation
during the 2003 French heat wave. While phones were extensively used between improvisers,
email use was associated with scarce improvisation. Furthermore, when improvisers involved
other responders into the ongoing improvisation, they were unsuccessful when they contacted
them by email or faxes. However new communication practices around faxes emerged during
the crisis response, which suggests that crisis responders can adapt their use of ICT tools
during crisis improvisation.

  273
 Figure 59. The final model

3.III. Final findings

In this section, we put into perspective theoretical and empirical findings.
Even though some ICT properties have a positive impact on the occurrence of organizational
improvisation constituents in crisis response, the effect of technology on organizational
improvisation growth is impacted by the communicative genres that have developed within
the responding organizations. As our empirical findings suggest, the 2003 French heat wave
crisis responders had developed what we label a “dispassionate” communicate genre around
emails. On the basis of our empirical findings, we discuss our theoretical propositions on ICT
properties’ influence on crisis improvisation. We consequently refine our findings on ICT
support to crisis organizational improvisation. We suggest that communicative genres develop
on both ICT properties and organizational and professional values. Thus communicative
genres mediate the effect of ICT on organizational improvisation in crisis response. We
develop this point in the following paragraphs.
First of all, communicative practices can distort the benefit of data calculation for crisis
improvisers over time. Data calculation refers to the possibility to treat a significant mass of
data to support decision making. Data calculation compels users to collect and transmit data
and figures. Our analysis of the 2003 French heat wave suggests the importance of figures in
the dispassionate communicative genre. As the administrative actors primarily exchanged
figures through electronic communication, the expression of their personal emotions and
opinions was depreciated in comparison with objective, measurable thus legitimate data. Even
though measures are critical to enact emergency procedures, they are not the only type of
reliable data. The administrative actors developed the common belief that figures were the
most legitimate data to transmit. As a result, figures became a predominant content within
electronic and email users developed data hunting practices.

  274
Secondly, virtuality promotes pioneering new communicative channels but can impede the
fast-evolution of existing communicative genres. Virtuality refers to the manipulation of data
and information as immaterial inputs. Thus ICT users can easily copy, cut or retrieve
information. Combined with modularity, virtuality promotes the quick settling of new
communication channels and minimal structures to regulate communication. For example,
Google Maps was spontaneously used to exchange information about damaged locations and
victims in a harmonized format. However, virtuality implies that individuals are likely to
develop rigidities in relation to ICT use. Contrasting with material media, e.g. that does not
allow data copy and retrieving without manipulating objects; virtuality prevents ICT users
from altering or coupling data with inputs that emerge on material content. On the contrary,
material media offer the opportunity to leave traces that are not necessarily sorted or ranged
with transmitted data, such as Post Its, written notes or manuscript comments on printed
reports. As a result, material media allows crisis responders to use other types of data than the
ones previously used. Virtuality can impede the adaptation of communicative genres when
users need a routine ICT tool, at the expense of the development of organizational
improvisation.
Thirdly, virtuality and many-to-many communication can result in their users developing
strategies to protect themselves from superiors and colleagues’ criticism. Forwarding an email
to collaborators only requires a few clicks, which makes it easy to report one’s activity to
colleagues and superiors, thereby mitigating the risk of being criticized for information
retention. In the summer of 2003, the heat wave crisis responders developed these strategies
to show off their own transparency through email use. However, the dispassionate
communicative genre influenced the crisis responders to the point that showing off one’s
transparency became the predominant objective in electronic communication. In addition, the
dispassionate communicative genre prevented crisis responders to discuss their emotions and
personal opinions.
What we call narration process can develop from data centralization. Data centralization
offers the possibility of easily retrieving past communication and supports organizational
improvisation by enabling crisis responders to jump on the bandwagon. Not only considered
to be an opportunity, data centralization can also be perceived as a threat. In the case of the
2003 French heat wave, the administrative actors had assimilated data centralization as a risk
of being caught for not respecting rules and procedures. To protect themselves from this risk
they primarily reported the tasks they completed in emails, limiting opportunities to discuss
ideas about what needed to be done.
Finally, our empirical investigation does not provide evidence of side effects from the
dispassionate communicative genre on modularity and graphical representation as ICT
properties. In Figure 60, some of the theoretical and empirical findings are represented and
connected by bold arrows. In this figure, the thin arrows correspond to the development of the
dispassionate communicative genre over the ICT properties that are likely to support
organizational improvisation in crisis response.

  275
 Figure 60. ICT properties and communicative genre effects on organizational
improvisation in crisis response

Conclusion

To conclude, the overall model of the dissertation presents a nuanced view of technology
support to organizational improvisation in crisis response. ICT properties offer the possibility
for crisis responders to improvise collectively. Additionally, ICT users develop
communicative genres on the basis of their professional and organizational values. Our
empirical investigation suggested that the 2003 French heat wave crisis responders perceived
electronic communication as a source of risk. As a result, email use was associated with the
objective of enhancing transparency by reporting completed or ongoing processes. Similarly,
data calculation morphed into data hunting practices. Such communication practices had a
significant influence on improvisation. Finally, crisis responders hardly moved from the
dispassionate communicative genre, which is consistent with theories on rigidity threats (Staw
et al., 1981).
To conclude, ICT properties are not only features of the technology artifact. They represent
opportunities in terms of coordination and control over organizational processes. Over their

  276
technical features, ICT properties have an organizational dimension as well. Our findings are
consistent with existing academic work on users’ enactement of technology (DeSanctis,
Poole, 1994; Orlikowski, 2000; Robey, Sahay, 1996). Table 46 presents an overview of the
dissertation research questions and findings.

Table 46. The dissertation research questions and findings

Research question Findings

1. How to cope with
diversity in research on
improvisation?
2. What possibilities does
ICT offer for interaction
during organizational
improvisation in crisis
response?
3. Does ICT fit users’ needs
with respect to interaction
during crisis improvisation?
4. How do users benefit (or
not) from ICT when they
interact to improvise
collectively in crisis
response?
Diversity originates from divergences between authors with
respect to four tasks that compose the research process on
improvisation: evidencing, recognizing, Contextualizing and
creating knowledge on improvisation.
We recommend that researchers:
1) Identify their perspectives on improvisation with respect to
these four tasks.
2) Identify the similarities between their approach and others
authors’ approach of improvisation to detect opportunities to
discuss their findings with similar research
We identify 5 constituents of organizational improvisation
We identify 6 ICT properties that support organizational
improvisation.
However the impact of the ICT properties is mediated by the
development of communicative genres on these properties and
organizational structure and culture.
Given that crisis response involves numerous individuals and
diverse organizations, communication tools are likely not to fit
to all crisis responders.
However, the theories of fit are not applicable to the specific
case of crisis improvisation without conceptual enrichment.
Our empirical findings suggest that the predominance of the
dispassionate communicative genre among email & fax users
prevented them from participating in crisis improvisation.
Conversely the fervent communicative genre, developed
around phone and face-to-face interactions, supported
organizational improvisation during the 2003 FHW.
Crisis improvisers’ benefit from ICT depends on their
communicative genres.

  277
Chapter 4.2 Conclusion

Introduction

In the dissertation project we attempt to shed light on ICT support to organizational

improvisation in crisis response. The relevance of our work is a core point of our discussion,
all the more as we embrace pragmatist principles. According to Lyytinen (1999b), the
relevance is “about what the researcher sees as practice and what elements are relevant in
understanding and changing practice”(p. 25), which makes it important to reflect what is
relevant for manager’s practice. For this reason, we strain to respond to the managerial needs
and the theoretical challenges. For more detail on these points, please refer to chapter 1.1.
Section 1 discusses our managerial and theoretical contribution. In section 2 we identify some
limitations of our work and propose some directions for future research to respond to these
limitations.

Section 1. Managerial contribution

Our study confirms that improvisation consists of a potentially relevant collective practice to
cope with crises. For this reason, we suggest guidelines for managers to recognize and
support ongoing improvisation in crisis response. However, it is important for managers not
to idealize improvisation as a magical solution to crisis response. Even though we overlook
the side effects of improvisation in this work, improvisation can be a disturbing experience
for both victims and crisis responders. During the 2003 heat wave, improvisation enabled
actors to find solutions to hydrate and care for many of those affected and ultimately save
lives. Undoubtedly, improvisation promoted a more efficient crisis response. At the same
time, administrative actors who improvised went under a lot of criticism from the public when
they used food vans to stock dead bodies, even though improvisation was eventually
perceived as necessary and even applauded.
1.I. Improvisation management
We recommend that managers and workers become more familiar with improvisation. This
statement is consistent with previous academic work that calls for a better understanding and
recognition of improvisation by managers (Barrett, 1998; Ciborra, 1996a; Ciborra et al.,
1999; Vera, Crossan, 2005, 2007). Some authors recommended the development of a trial-
and-error approach to problem solving (Eisenhardt, Tabrizi, 1995). Other authors rely on
theatre and music improvisation as means to develop personal improvisational capabilities
(Shaw, Stacey, 2006). We expand on these recommendations by suggesting that managers
and workers can learn to identify episodes of improvisation. With respect to crisis
improvisation, we suggest that workers can use analysis grids to understand their own
improvisation. Figure 61 provides a fictional example of an analysis grid of improvisation
that we developed from our findings and that we label “Organizational improvisation
roadmap”. This figure represents a form that was completed by a nurse who had to improvise
an ad hoc solution to respond to the power failure of a building air conditioner in the midst of

  278
a hot summer. We suggest that workers fill the improvisational roadmap forms in the
aftermaths of crisis response and have the option to meet other improvising colleagues to
share their thoughts on their roadmaps. By doing so, they can enrich a repertory of possible
improvisation, thus realizing that improvisation is a part of their daily practices and a support
to the emergence of new capabilities, as we suggest in chapter 2.3. In addition, such a
collective reflection on improvisation can help managers to promote improvisation within
organizations.

  279
 Figure 61. A fictional example of improvisation roadmap

  280
 1.II.ICT and communicative genres
Our investigation provides some evidence that ICT investments are worthwhile to support
organizational improvisation. However, managers have to handle potential threats to an
efficient use of ICT tools when improvisation is needed. More specifically they have to
address the potential side effects of user’s appropriation of the ICT properties. For instance,
they have to prevent email users from overlooking other type of information than measurable
data. More generally, they have to detect whether communicative genres burden
organizational improvisation and to promote the development of an improvisation friendly
communicative genre. Previous academic work provides analysis grids as means to identify
communicative genres (Orlikowski, Yates, 1994; Yates, Orlikowski, 2002)
Our empirical investigation suggests that professional and organizational values play an
important role in the development of communicative genres. For instance, showing
evidencing that one’s actions abide by the rules and the procedures is a well-known cultural
specificity in the French health care organizations (Morgan, Krone, 2001). There is no
surprise then that actors developed a dispassionate communicative genre whose objective is to
provide evidence of rules being applied. However, it is important to provide a more nuanced
picture of the dispassionate communicative genre: we believe that the dispassionate
communicative genre does not have only negative effects on crisis improvisation.
Accordingly, we recommend that managers try to support diversity within communicative
genres. Our suggestions focus on the management of communicative genres related to ICT
use.
First of all we propose that organizations create virtual spaces devoted to informal and
discursive interactions. For instance, a specific space on the Intranet can be opened to every
organizational member for brainstorming, polls, open discussion on controversial topics and
informal sharing about professional experiences. We recommend that managers grant that this
space is free of control and arbitrary moderation thanks to an implicit contract or explicit
chart with users. Reciprocally, users should engage of using it in a constructive manner,
thereby avoiding the mentioning of rumors or vilification. In our view, this virtual space is a
possible starting point to promote the fervent communicative genre via electronic
communication. However, our suggestion implies security issues. It contradicts Turoff and his
colleagues’ recommendations (Turoff et al., 2004) to rigorously define the users’ rights in
relation electronic communication. For this reason, we suggest that the use of a virtual space
remain secondary to primarily used communication media. By their existence, free virtual
spaces can reassure users that they can interact in a fervent genre.
Secondly, we suggest that the use of diverse media for communication is valuable in crisis
response. Our investigation suggests that users adapted their communicative genre more or
less easily, depending on the media used. If we take a further look at faxes and email use
during the 2003 French heat wave, we can find that crisis responders adapted their
communicative genres to a more fervent genre by writing on faxes. When mortuary chambers
became filled with bodies, administrative directors at the head of the health care national an

  281
regional organizations decided to rent food vans to stock the corpses. They used faxes to
communicate directions to hospitals directors on how to transfer dead bodies to food vans
they had rented. At that time, they adjusted their manner of expression by writing additional
comments by hand on fax letters. Our suggestion contradicts Jennex’s recommendation to
restrict communication to a unique and secure media (2008). We do agree that
communicating through several media can imply the risk of dispersion, misunderstandings,
and information overload (Hiltz, Turoff, 1985; Jennex, 2008; Turoff et al., 2004). However,
the use of multiple media reduces the odds that improvisation will be blocked by
communicative genres in crisis response.
As a conclusion to this section, Table 47 presents the managerial needs that we address in this
study. For a more complete description of the managerial needs related to our research project
please refer to chapter 1.1

  282
 Table 47.Managerial needs addressed in the dissertation

Key Issue Managerial needs Response

Crises become more The need to multiply potential Improvisation is a potentially relevant
and more frequent techniques and resources to technique to face crisis. However, it
(Section 1) face them (Perrow, 2006): requires care management because
improvisation does not necessarily
develop to the highest level,
depending on crisis responders’
communication practices

Technology is a crucial The need to evaluate the Technological investments are

means for coordination relevance of technological relevant but their efficiency depends
in crisis response but is investment to prepare for crisis on user’s practices primarily.
also likely to burden response
crisis response or at
least reduce The need to develop practices Manage free virtual spaces to promote
organizational to properly manage ICT and alternative communicative genres
resilience (Section 2) avoid technological pitfalls Use two or three communication
during crisis response media in crisis response

Managers remain Managers need some concrete Crisis responders and managers can
skeptical about it and guidelines about what use improvisation roadmaps to
theory develops slowly improvisation is or is not, how identify organizational improvisation.
(Section 3). to figure and manage it

There is no evidence of Managers are reluctant to Improvisation potentially enables

improvisation being mindfully monitor important resources leveraging and
beneficial to crisis improvisation in crisis innovative response to crisis response
response (Section 4). response.

ICT support to Managers need good practices Managers can promote:

improvisation is not to manage tensions between 1) the development of a fervent
established. (Section 5) improvisation requirements communicative genre
and ICT infrastructure. 2) the adaptability of existing
communicative genres

Section 2. Theoretical Contribution

Our theoretical contribution comprises the following points that are presented in three studies:
- Providing an integrative view of the sources of diversity with respect to the concept of
improvisation (chapter 2.2)

  283
- Proposing a definition of organizational improvisation. In particular, we address the
lack of knowledge on organizational improvisation by clarifying the role of interactions in
this process (chapter 2.3).
- Suggesting some technology characteristics that could support organizational
improvisation in crisis response (chapter 2.3).
- Understanding ICT use practices and their effects on organizational improvisation
during the 2003 French heat wave crisis (chapter 3.3).
- Providing further insights on the interactive dimension of crisis improvisation by
suggesting that communicative genre can leverage or inhibit improvisation during crisis
response. By doing so, we hope to provide new insights on technology use in crisis response,
which has been an overlooked theme in the MIS field so far (chapter 3.3) As a conclusion to
this section, Table 48 presents the theoretical challenges that we address in this work.

Table 48. Theoretical challenges addressed in the dissertation

Key Issue Theoretical challenges Responses

Crises become more The need to develop our We made a theoretical proposition
and more frequent knowledge on some unexplored on crisis organizational
(Section 1) aspects of crisis response such as improvisation
improvisation.
Technology is likely The need to report what makes Our findings suggest that
to burden crisis ICT become an obstacle to crisis technology is not an obstacle to
response or at least response improvisation but that
reduce communication practices that
organizational involve ICT tools can hamper
resilience (Section organizational improvisation
2)
Managers remain There is the challenge to some We make some theoretical
skeptical about it inner contradictions to the propositions on the constituents of
and theory develops concept of improvisation. organizational improvisation in
slowly (Section 3). There is the need to develop an chapter 2.3.
integrative view of interactions
that occur during organizational
improvisation

  284
There is no evidence There is little explanation of the Our preliminary analysis of the
of improvisation occurrence of improvisation in management literature on
being beneficial to critical settings improvisation suggests that
crisis response improvisation can endogenously
(Section 4). develop over time and/or be
triggered by an exogenous event
such as in crisis response.
Our empirical findings suggest
that organizational improvisation
development depends on
communicative genres.

ICT support to Understanding how people use We provide an integrative

improvisation is not ICT when they improvise explanation of organizational
established. (Section How does ICT use affect improvisation and ICT use during
5 + Section 6) improvisation? crisis response
The need to better understand
the nature of ICT use during
organizational improvisation

Section 3. Limitations and avenues for future research

3.I. Improvisation
This work only provides limited insights on improvisation. We propose here three possible
avenues for further investigation of this topic.
First, additional observation is required on the specificity of improvisation. On one hand, our
empirical investigation evidences improvisation as a federating process over organizations.
On the other, many authors highlight that improvisation is a situated performance, deeply
influenced by local cultural values (Aram, Walochik, 1996; Ciborra, Willcocks, 2006).
Further investigation is therefore required to clarify the role of organizational and professional
values in the occurrence and development of organizational improvisation. This suggestion
brings us back to one of the directions for research that we propose in chapter 2.2 and
introduce as the following question: Does the concept of improvisation refer to a unique
phenomenon or a diversity of phenomena? A possible direction for research could be to adopt
an ethnographic design to investigate the different cultural or professional contexts in which
improvisation occurs, such as NPD or electronic commerce.
Second, further reflection is required to distinguish effective improvisation from ineffective
improvisation. In the previous section, we recommend that managers promote improvisation
within their organizations. To promote improvisation, managers have to anticipate the
possible side effects of collateral damages due to failed improvisation. However,

  285
organizations lack guidelines to recognize and handle failing improvisation. Further
investigation is necessary to find cases of failed improvisation and to identify possible
consequences of ineffective improvisation. What is a failed improvisation? What is a
successful improvisation?
The third avenue of research related to improvisation addresses the effect of organizational
improvisation on organizational structures and culture. We have evidenced some of the
constituents of organizational improvisation but we fall short of details on the implications of
improvisation on organizational functioning.
3.II. Communicative genres
In this dissertation, we explore the effect of communicative genres on crisis response and
crisis improvisation. However, further understanding of the role played by communicative
genres in organizations is needed. Our empirical investigation focuses on the effect of
communicative genres on organizational improvisation. It identifies two communicative
genres that developed around several types of media. During the 2003 French heat wave, the
fervent communicative genre supported organizational improvisation. On the contrary, the
dispassionate communicative genre led crisis responders to avoid mentioning emotions and
opinions in their emails and faxes, even if emotions and opinions were important ingredients
of improvisation. Our findings depict the potential negative effect of the dispassionate
communicative genre on improvisation. However, one can easily find some positive effects of
the dispassionate communicative genre. In particular, the dispassionate communicative genre
can help crisis responders to keep distance with their own emotions during crisis response.
Even though emotions can leverage personal involvement in crisis response, they are likely to
disturb crisis responders, alter their judgment or even hinder their focus in decision-making.
Some managers even make desperate decision making under the effect of panic (Starbuck et
al., 1978). In addition, keeping crisis responders focused on a specific topic, the dispassionate
communicative genre can prevent crisis responders from getting lost in information overload
and help them to prioritize. Thus we can formulate various hypotheses in relation to the
combination of dispassionate and fervent communicative genres in organizations. We could
hypothesize that too much of the fervent communicative genre can impede organizational
improvisation (cases A, B and C on Figure 62). An alternate hypothesis could be that a certain
amount of fervent interactions are required to support organizational improvisation (case C on
Figure 62). Further investigation is required to clarify the effect of the combination of
multiple communicative genres on crisis improvisation.

  286
 Figure 62. Proportion of fervent and dispassionate communicative genres

3.III. Findings generalizability

The generalizability of our findings can be questioned. At first glance, our findings may not
be relevant for improvisation in general. Indeed, our analysis is contextualized and
idiographic (Mason et al., 1997 ). Moreover, it concerns organizations from health care
professions and is retrospective. In that sense, one could argue that it provides too little
material to enable companies to handle future challenges. As we explain in the empirical part
of the dissertation, the response network included a diverse panel of organizations, ranging
from hospitals to administration and private companies. Thus our results are not restricted to a
specific type of process or organization.
With respect to the retrospective dimension of our work, we agree with Marrou (1954) that
knowledge about our past is a major component of our knowledge about our present.
Knowing what people did in the past is useful to avoid mistakes in the coming years (Prost,
1996). In addition, we rely on grounded theory process of abstraction to develop extemporal
findings, such as the concept of organizational emptiness. In that sense, we view our findings
valuable for the present and future of companies. However, the issue of generalizability
remains in that there is no warrantee that we can generalize our results to other types of crises
than natural disasters provoked by natural hazards. Further investigation on other cases of
crisis improvisation would permit comparative approaches to generalize our findings.
3.IV. Improvisation and dynamic capabilities
Fourthly, the conceptual connections between improvisation and dynamic capabilities remain
vague. We draw inspiration from the dynamic capabilities approach to make theoretical

  287
propositions on organizational improvisation (for more detail, see chapter 2.3). Some studies
provided empirical support to our theoretical propositions on improvisation (Adrot, Garreau,
2010), suggesting that the dynamic capabilities approach is relevant for a further
understanding of organizational improvisation as a collective process.
The dynamic capabilities approach represents one of the prominent theories in strategic
management and does not originally deal with crisis response. Dynamic capabilities enable
organizations to develop new capabilities in a short delay to maintain a competitive
advantage, which has little in common with an organization’s effort to survive and face a
disruptive unexpected event. As suggested by Wang and Ahmed’s definition, dynamic
capabilities can be conceptualized as an organizations’ “behavioral orientation constantly to
integrate, reconfigure, renew and recreate its resources and capabilities, and most
importantly, upgrade and reconstruct its core capabilities in response to the changing
environments to attain and sustain competitive advantage” (Wang, Ahmed, 2007).
However, we argue that the ongoing academic work on routine and dynamic capabilities
provides valuable insights on organizational improvisation during crisis response. Consider
the ad hoc solutions that emerge from a local improvisation. When improvisation grows
organizational, these ad hoc solutions are discussed, repeated and integrated within the
organization. In addition, crisis improvisers develop new coordination paths to collectively
implement these solutions. For example, during the 2003 French heat wave, some
administrative actors adapted their use of faxes to transmit directions in relation to the
improvised transfer of corpses. Similarly, hospitals developed new rules to manage the flow
of patients. In our view, collective adaptation to implement improvisation outputs corresponds
to Zollo and Winter’s findings on dynamic capabilities (2002). Further investigation is
required to clarify the conceptual ties between improvisation and dynamic capabilities. More
specifically, future research could address the two following questions: Is improvisation a
dynamic capability? Do dynamic capabilities support crisis improvisation?
3.V. Organizational emptiness
Fifthly, the concept of organizational emptiness emerged from our analysis of empirical data.
As we explain in chapter 3.3, numerous authors refer to organizational emptiness more or less
explicitly. Research regularly mentions lack, gaps, needs and even the notion of
organizational holes, blanks. Clarifying the conceptual ties between organizational emptiness
and these concepts is necessary.
Finally, we suggest that organizational emptiness represents an opportunity for improvisation
in chapter 3.3. However, we lack knowledge on the role that organizational emptiness play in
organizational functioning. In particular, further research could address the following
question: Is emptiness a good thing or a bad thing for organizations?

  288
  289
 References
ADROT A, and GARREAU L (2010) Interagir pour improviser en situation de crise. Le cas
de la canicule de 2003. Revue française de gestion 36 (203), pp. 119-131.
AHITUV N, IGBARIA M, and SELLA A (1998) The effects of time pressure and
completeness of information on decision making. Journal of management information
systems 15 (2), pp. 153-172.
AKGÜN AE, and LYNN GS (2002) New product development team improvisation and
speed-to-market: An extended model. European Journal of Innovation Management 5
(3), pp. 117-119.
ALAVI M, and LEIDNER DE (2001) Review: Knowledge management and knowledge
management systems: Conceptual foundations and research issues. MIS Quarterly 25
(1), pp. 107-136.
ALDENDERFER MS, and BLASHFIELD RK (1984) Cluster analysis. Sage, Beverly Hills,
CA.
ALTER N (2000) L'innovation ordinaire. PUF, Paris.
AMMENWERTH E, BUCHAUER A, BLUDAU B, and HAUX R (2000) Mobile
information and communication tools in the hospital. International Journal of Medical
Informatics 57 pp. 21-40.
ANDERSON RA, and MCDANIEL RRJ (2000) Managing health care organizations: Where
professionalism meets complexity science. Health Care Management Review 25 (1),
pp. 83-92.
ANDERSON RC, and ORTONY A (1975) On putting apples into bottles - a problem of
polysemy. Cognitive Psychology 7 pp. 167-180.
ARAM JD, and WALOCHIK K (1996) Improvisation and the spanish manager. International
studies of management & organization 26 (4), pp. 73-89.
ARGYRIS C, and SCHON D (1974) Theory in practice: Increasing professional
effectiveness. Jossey-Bass, Oxford, England.224 p.
BAKER T (2007) Resources in play: Bricolage in the toy store (y). Journal of Business
Venturing 22 (5), pp. 694-711.
BAKER T, MINER AS, and EESLEY DT (2003) Improvising firms: Bricolage, account
giving and improvisational competencies in the founding process. Research Policy 32
(2), pp. 255-276.
BAKER T, and NELSON RE (2005) Creating something from nothing: Resource
construction through entrepreneurial bricolage. Administrative Science Quarterly 50
(3), pp. 329-366.
BALES RF (1950) A set of categories for the analysis of small group interaction. American
Sociological Review 15 (2), pp. 257-263.
BAMBERGER PA, and PRATT MG (2010) Moving forward by looking back: Reclaiming
unconventional research contexts and samples in organizational scholarship. Academy
of Management Journal 53 (4), pp. 665-671.

  290
BANDEIRA-DE-MELLO R, and GARREAU L (2008) Possibilités et pièges liés à
l'utilisation des logiciels dans le processus d'analyse au travers de la théorie enracinée.
In AIMSEds).
BANSLER J, and HAVN E (2003) Improvisation in action: Making sense of is development
in organizations. In The International Workshop on Action in Language, Organization
and Information Systems (ALOIS)Eds).
BANSLER J, and HAVN E (2004) Improvisation in information systems development. In:
Information systems researchEds.), pp. 631-646,
BARKHI R, and SHEETZ SD (2001) The stats of theoretical diversity. Communications of
the Association for Information Systems 7 (6),
BARLEY SR (1986) Technology as an occasion for structuring: Evidence from observations
of ct scanners and the social order of radiology departments. Administrative Science
Quarterly 31 (1), pp. 78-108.
BARNETT CK, and PRATT MG (2000) From threat-rigidity to flexibility - toward a learning
model of autogenic crisis in organizations. Journal of Organizational Change
Management 13 (1), pp. 74-88.
BARNEY J (1991) Firm resources and sustained competitive advantage. Journal of
Management 17 (1), pp. 99.
BARRETT FJ (1998) Creativity and improvisation in jazz and organizations: Implications for
organizational learning. Organization Science 9 (5), pp. 605-622.
BARRY D, and RERUP C (2006) Going mobile: Aesthetic design considerations from calder
and the constructivists. Organization Science 17 (2), pp. 262-276.
BARTON AH (1969) Communities in disaster. Anchor, Doubleday, Garden City, NY.352 p.
BASTIANELLI A-P, FOURCADE M, PICARD S, and SALZBERG L "Mission d'enquête
sur les fermetures de lits en milieu hospitalier durant l'été 2003," Inspection Générale
des Affaires Sociales, Paris, p. 52.
BATESON G (1935) Culture contact and schismogenesis. Management Science 35 pp. 178-
183.
BEACH LR, and MITCHELL TR (1978) A contingency model for the selection of decision
strategies. The Academy of Management review 3 (3), pp. 439-449.
BEAMISH TD (2002) Silent spill. The MIT Press,
BENBASAT I, and WEBER R (1996) Research commentary: Rethinking "Diversity" In
information systems research. Information Systems Research 7 (4), pp. 389-399.
BENBASAT I, and ZMUD RW (1999) Empirical research in information systems: The
practice of relevance. MIS Quarterly 23 (1), pp. 3-16.
BENSON JK (1977) Organizations: A dialectical view. Administrative Science Quarterly 22
(1), pp. 1.
BIGLEY GA, and ROBERTS KH (2001) The incident command system: High-reliability
organizing for complex and volatile task environments. Academy of Management
Journal 44 (6), pp. 1281-1300.

  291
BILLINGS RS, MILBURN TW, and SCHAALMAN ML (1980) A model of crisis
perception: A theoretical and empirical analysis. Administrative Science Quarterly 25
(2), pp. 300-316.
BLAIKIE N (2007) Approaches to social enquiry. Polity, 234 p.
BLUMER H (1986) Symbolic interactionism: Perspective and method. University of
california Press, Berkeley.208 p.
BOIN A (2004) Lessons from crisis research. International studies review 6 (1), pp. 165-194.
BOIN A, and LAGADEC P (2000) Preparing for the future: Critical challenges in crisis
management. Journal of contingencies and crisis management 8 (4), pp. 185-191.
BOUDES T, and LAROCHE H (2009) Taking off the heat: Narrative sensemaking in post-
crisis inquiry reports. Organization Studies pp. 377-396.
BOUKEF N "Utilisation du courrier électronique dans l'activité managériale: Usages, intérêts
et limites," in: CREPA, Paris-Dauphine University, Paris, 2005, p. 325.
BOURDIEU P (1980) Le sens pratique. Les Editions de Minuit, Paris.
BOWKER GC, and STAR SL (2000) Sorting things out: Classification and its consequences.
MIT Press, Cambridge, Massachusetts.377 p.
BOYER L, ROBITAIL S, DEBENSASON D, AUQUIER P, and SAN MARCO JL (2005)
Média et santé publique : L'exemple de la canicule pendant l'été 2003 en france. Revue
d'Épidémiologie et de Santé Publique 53 (5), pp. 525-534.
BRANDON DP, and HOLLINGSHEAD AB (2004) Transactive memory systems in
organizations: Matching tasks, expertise, and people. Organization Science 15 (6), pp.
633-644.
BRIGHAM M, and INTRONA LD (2006) Hospitality, improvisation and gestell: A
phenomenology of mobile information. Journal of Information Technology 21 (3), pp.
140-153.
BROOKS RG, and MENACHEMI N (2006) Physicians' use of email with patients: Factors
influencing electronic communication and adherence to best practices. Journal of
Medical Internet Research 8 (1),
BROWN JS, and DUGUID P (1991) Organizational learning and communities-of-practice:
Toward a unified view of working, learning, and innovation. Organization Science 2
(1), pp. 40-57.
BROWN SL, and EISENHARDT KM (1997) The art of continuous change: Linking
complexity theory and time-paced evolution in relentlessly shifting organizations.
Administrative Science Quarterly 42 (1), pp. 1-34.
BURRELL G, and MORGAN G (1979) Sociological paradigms and organisational analysis.
Heinemann Educational Books, London.
BURT RS (1992) Structural holes: The social structure of competition. Harvard University
Press, Cambridge, MA.
BURT RS (2004) Structural holes and good ideas. The American Journal of Sociology 110
(2), pp. 349-399.

  292
CAI G, MACEACHREN AM, BREWER I, MCNEESE M, SHARMA R, and FUHRMANN
S (2005) Map-mediated geocollaborative crisis management. In ISI 2005, (KANTOR
P, Eds).
CALLOWAY LJ, and KEEN PGW (1996) Organizing for crisis response. Journal of
Information Technology 11 (1), pp. 13-26.
CAMERON KS, WHETTEN DA, and KIM MU (1987) Research notes organizational
dysfunctions of decline. Academy of Management Journal 30 (1), pp. 126-138.
CAPLAN G (1964) Principles of preventive psychiatry. Basic Books, New York.
CAR J, and SHEIKH A (2004) Email consultations in health care: 2--acceptability and safe
application. Bmj 329 (7463), pp. 439-442.
CARREL LF (2005) Epidemic in switzerland”: Description of a strategic leadership exercise
by the swiss government. Journal of contingencies and crisis management 13 (4), pp.
170-175.
CARVER L, and TUROFF M (2007) Human-computer interaction: The human and computer
as a team in emergency management information systems. Communications of the
ACM 50 (3), pp. 33-38.
CHAN TC, KILLEEN J, GRISWOLD W, and LENERT L (2004) Information technology
and emergency medical care during disasters. Academic emergency medicine 11 (11),
pp. 1229-1236.
CHARMAZ K (2006) Constructing grounded theory: A practical guide through qualitative
analysis. Sage, 208 p.
CHELARIU C, JOHNSTON WJ, and YOUNG L (2002) Learning to improvise, improvising
to learn - a process of responding to complex environments. Journal of Business
Research 55 (2), pp. 141-147.
CHEN Y, BRENNAN N, and MAGRABI F (2010) Is email an effective method for hospital
discharge communication? A randomized controlled trial to examine delivery of
computer-generated discharge summaries by email, fax, post and patient hand
delivery. International Journal of Medical Informatics 79 (3), pp. 167-172.
CHERRYHOLMES CH (1992) Notes on pragmatism and scientific realism. Educational
Researcher 21 (6), pp. 13-17.
CHUA AY (2007) A tale of two hurricanes: Comparing katrina and rita through a knowledge
management perspective. Journal of the American Society for Information Science
and Technology 58 (10), pp. 1518-1528.
CIBORRA C (1996a) Improvisation and information technology proceedings in
organizations. In ICIS Eds).
CIBORRA C (1996b) The labyrinths of information: Challenging the wisdom of systems.
Oxford University Press, Oxford.
CIBORRA C, BRAA K, and CORDELLA A (1999) From control to drift. Oxford University
Press,
CIBORRA C, and HANSETH O (1998) From tool to gestell: Agendas for managing
information infrastructure. Information, Technology & People 11 (4), pp. 305-327.
CIBORRA C, and WILLCOCKS L (2006) The mind or the heart? It depends on the
(definition of) situation. Journal of Information Technology 21 pp. 129-139.

  293
CIBORRA CU (1999) Notes on improvisation and time in organizations. Accounting,
management, and information technologies 9 (2), pp. 77-94.
CLAIR JA, and DUFRESNE RL (2005) Phoenix rising - positive consequences arising from
organizational crisis. In: Positive psychology in business ethics and corporate
responsibility - ethics and the environment, (GIACALONE RA, JURKIEWICZ CL
and DUNN C, Eds.), IAP
CLEGG S (2002) Management and organization paradoxes.
CLEGG SR, CUNHA JV, and CUNHA MPE (2002) Management paradoxes: A relational
view. Human Relations 55 (5), pp. 483-503.
COHEN MD (1994) Organizational routines are stored as procedural memory: Evidence from
a laboratory study. Organization Science 5 (4), pp. 554.
COHEN MD, MARCH JG, and OLSEN JP (1972) A garbage can model of organizational
choice. Administrative Science Quarterly 17 (1), pp. 1-25.
COLLINS LM, POWELL JE, DUNFORD CE, MANE KK, and MARTINEZ MLB (Year)
Emergency information synthesis and awareness using e-sos. In 5th International
ISCRAM ConferenceEds)Washington, DS, USA.
COMFORT L, DUNN M, JOHNSON D, SKERTICH R, and ZAGORECKI A (2004)
Coordination in complex systems: Increasing efficiency in disaster mitigation and
response. International Journal of Emergency Management 2 (1-2), pp. 62-80.
COMFORT LK (1993) Integrating information technology into international crisis
management and policy. Journal of contingencies and crisis management 1 (1), pp.
15-26.
COMFORT LK (2007) Crisis management in hindsight: Cognition, communication,
coordination, and control. Public Administration Review 67 pp. 189-197.
COMFORT LK, SUNGU Y, JOHNSON D, and DUNN M (2001) Complex systems in crisis:
Anticipation and resilience in dynamic environments. Journal of contingencies and
crisis management 9 (3), pp. 144-158.
CORBIN JM, and STRAUSS A (1990) Grounded theory research: Procedures, canons, and
evaluative criteria. Qualitative Sociology 13 (1), pp. 3-21.
COURBON J-C, and TAJAN S (1997) Groupware et intranet. InterEditions, 230 p.
COWTON CJ (1998) The use of secondary data in business ehtics research. Journal of
Business Ethics 17 pp. 423-434.
CROSSAN M, CUNHA MPE, VERA D, and CUNHA J (2005) Time and organizational
improvisation. Academy of Management Review 30 (1), pp. 129-145.
CROSSAN MM (1998) Improvisation in action. Organization Science 9 (5), pp. 593-599.
CROSSAN MM, LANE HW, WHITE RE, and KLUS L (1996) The improvising
organization: Where planning meets opportunity. Organizational Dynamics 24 (4), pp.
16.
CROSSAN MM, and SORRENTI M (2002) Making sense of improvisation. In:
Organizational improvisation, (KAMOCHE K, CUNHA MP and CUNHA J, Eds.),
pp. 27, Routledge.
CROZIER M, and FRIEDBERG E (1977) L'acteur et le système. Seuil, Paris.

  294
CUNHA JV (2001) Brave new(paradoxical) world: Structure and improvisation in virtual
teams. Strategic change 10 (6), pp. 337.
CUNHA MP "Management improvisation," in: FEUNL Working paper, FEUNL, 2004.
CUNHA MP, CUNHA J, and FAIA-CORREIA M (1999a) "Scenarios for improvisation long
range planning deemed," in: Warwick Business School Research Papers, Warwick
Busness School, Warwick.
CUNHA MP, CUNHA JV, and KAMOCHE K (1999b) Organizational improvisation: What,
when, how and why. International journal of management reviews 1 (3), pp. 299.
CUNHA MP, CUNHA JV, and KAMOCHE K (1999c) Organizational improvisation: What,
when, how and why. International Journal of Management Reviews 1 (3), pp. 299-
341.
CUNHA MP, and GOMES JFS (2003) Order and disorder in product innovation models.
Creativity and innovation management 12 (3), pp. 174-187.
CUNHA PM, KAMOCHE K, and CUNHA RC (2003) Organizational improvisation and
leadership-a field study in two computer-mediated settings. International studies of
management & organization 33 (1), pp. 34-57.
DABBISH LA, KRAUT RE, FUSSELL SR, and KIESLER S "Understanding email use:
Predicting action on a message," in: Proceedings of the SIGCHI conference on Human
factors in computing systems, ACM, Portland, Oregon, USA, 2005.
DAFT RL, LENGEL RH, and TREVINO LK (1987) Message equivocality, media selection,
and manager performance: Implications for information systems. MIS Quarterly 11
(3), pp. 355-366.
DANTAS A, and SEVILLE E (2006) Organisational issues in implementing and information
sharing framework: Lessons from the matata flooding events in new zealand. Journal
of Contingencies & Crisis Management 14 (1), pp. 38-52.
DAVIS FD (1989) Perceived usefulness, perceived ease of use, and user acceptance of
information technology. MIS Quarterly 13 (3), pp. 319-340.
DAVIS J, EISENHARDT K, and BINGHAM C (2009) Optimal structure, market dynamism,
and the strategy of simple rules. Administrative Science Quarterly 54 (3), pp. 413.
DAVIS JB "The impact of natural disasters on employee turnover: The shocks and ater-
shocks of hurricane katrina in it professionals," in: Interdepartemental Program in
Business Administration, Louisiana State University, 2008, December, p. 137.
DAWES SS, CRESSWELL AM, and CAHAN BB (2004) Learning from crisis - lessons in
human and information infrastructure from the world trade center response. Social
Science Computer Review 22 (1), pp. 52-66.
DAYTON BW (2004) Managing crises in the twenty-first century. International studies
review 6 (1), pp. 29.
DE VAUS DA (2001) Research design in social research. Sage, 279 p.
DEANE PD (1988) Polysemy and cognition. Lingua 75 (4), pp. 325-361.
DEETZ S (1996) Describing differences in approaches to organization science: Rethinking
burrell and morgan and their legacy. Organization Science 7 (2), pp. 191-207.

  295
DENIS J-L, LAMOTHE L, and LANGLEY A (2001) The dynamics of collective leadership
and strategic change in pluralistic organizations. The Academy of Management
Journal 44 (4), pp. 809-837.
DENNIS AR, and KINNEY ST (1998) Testing media richness theory in the new media: The
effects of cues, feedback, and task equivocality. Information Systems Research 9 (3),
pp. 256-274.
DESANCTIS G, and POOLE MS (1994) Capturing the complexity in advanced technology
use: Adaptive structuration theory. Organization Science 5 (2), pp. 121-147.
DESTAIS N (2003) Le système de santé: Organisation et régulation. LGDJ, Paris.256 p.
DEWEY J (1908) What does pragmatism mean by practical? The journal of philosophy,
psychology and scientific methods 5 (4), pp. 85-99.
DRAZIN R, and VAN DE VEN AH (1985) Alternative forms of fit in contingency theory.
Administrative Science Quarterly 30 (4), pp. 514-539.
DUCHENEAUT N, and BELLOTTI V (2001) E-mail as habitat: An exploration of embedded
personal information management. interactions 8 (5), pp. 30-38.
DUCHENEAUT NB (2002) The social impacts of electronic mail in organizations: A case
study of electronic power games using communication genres. Information,
Communication & Society 5 (2), pp. 153 - 188.
DURRIEU F, and VALETTE-FLORENCE P (2005) L'analyse typologique: De l'exploratoire
au confirmatoire. In: Management des ressources humaines: Méthodes de recherche
en sciences humaines et sociales, (WACHEUX E and ROUSSEL P, Eds.), pp. 379-
429, Broché, Paris.
DYBA T (2000) Improvisation in small software organizations. IEEE Software 17 (5), pp. 82-
82.
DYNES R (1970) Organized behavior in disaster. Health Lexington Books, Mass,
DYNES R, and QUARANTELLI EL "Organizational communications and decision making
in crises," Ohio State University Columbus Disaster Research Center.
EDGLEY R (2003) Reason as dialectic: Science, social science and socialist science. In:
Critical realism: Essential readings, (ARCHER M, BASHKAR R, COLLIER A,
LAWSON T and NORRIE A, Eds.), pp. 395-408, Routledge
EDVARDSSON B, HAGLUND L, and MATTSSON J (1995) Analysis, planning,
improvisation and control in the development of new services. International Journal
of Service Industry Management 6 (2), pp. 24-&.
EISENBERG EM (2006) Karl weick and the aesthetics of contingency. Organization Studies
27 (11), pp. 1693.
EISENHARDT K, and TABRIZI BN (1995) Accelerating adaptive processes: Product
innovation in the global computer industry. Administrative Science Quarterly 40 (1),
pp. 84-110.
EISENHARDT KM (1989) Making fast strategic decisions in high-velocity environments.
The Academy of Management Journal 32 (3), pp. 543-576.
EISENHARDT KM (2000) Dynamic capabilities: What are they? Strategic management
journal 21 (10-11), pp. 1105.

  296
ESPINOSA A, LERCH J, and KRAUT R (2002) Explixit vs. Implicit coordination
mechanisms and task dependencies: One size does not fit all. In: Forthcoming in team
cognition: Process and performance at the inter- and intra-individual level, (SALAS
E, FIORE SM and CANNON-BOWERS JA, Eds.),
EVIN C "Rapport fait au nom de la commission d'enquête sur les conséquences sanitaires et
sociales de la canicule," Assemblée Nationale.
FAIA-CORREIA M (2003) Mind the gap between processes and practice!-the role of
improvisation in the implementation of information technology. International studies
of management & organization 33 (1), pp. 58.
FARAJ S, and SPROULL L (2000) Coordinating expertise in software development teams.
Management Science 46 (12), pp. 1554-1568.
FARAJ S, and XIAO Y (2006) Coordination in fast-response organizations. Management
science 52 (8), pp. 1155-1169.
FAUCHART E (2006) Moral hazard and the role of users in learning from accidents. Journal
of contingencies and crisis management 14 (2), pp. 97-106.
FELDMAN MS (2000) Organizational routines as a source of continuous change.
Organization Science 11 (6), pp. 611-629.
FELDMAN MS, and PENTLAND BT (2003) Reconceptualizing organizational routines as a
source of flexibility and change. Administrative Science Quarterly 48 (1), pp. 94-118.
FESTINGER L (1957) A theory of cognitive dissonance. Standford University Press, 291 p.
FICHMAN RG "Information technology diffusion: A review of empirical research," in:
Proceedings of the thirteenth international conference on Information systems,
University of Minnesota, Dallas, Texas, United States, 1992.
FIESCHI M "Les données du patient partagées: La culture du partage et de la qualité des
informations pour améliorer la qualité des soins," Ministère de la santé, de la famille
et des personnes handicapées, Paris, p. 56.
FINK SL, BEAK J, and TADDEO K (1971) Organizational crisis and change. The Journal of
applied behavioral science 7 (1), pp. 15-37.
FIOL CM, and LYLES MA (1985) Organizational learning. Academy of Management Review
10 (4), pp. 803-813.
FISCHER III HW (1996) What emergency management officials should know to enhance
mitigation and effective disaster response. Journal of Contingencies & Crisis
Management 4 (4), pp. 208.
FITRIANIE S, POPPE R, BUI TH, CHITU AG, DATCU D, DOR R, HOFS DHW,
WIGGERS P, WILLEMS DJM, POEL M, ROTHKRANTZ LJM, VUURPIJL LG,
and ZWIERS J (Year) A multimodal human-computer interaction framework for
research into crisis management. In ISCRAM, (VAN DE WALLE B, BURGHART P
and NIEUWENHUIS C, Eds).
FOR-MUKWAI GF (2010) The transformative power of social media on emergency and
crisis management. International Journal of Information Systems for Crisis Response
and Management 2 (1), pp. 1-10.
FORD R (2008) Complex adaptive systems and improvisation theory: Toward framing a
model to enable continuous change. Journal of Change Management 8 (3/4), pp. 173.

  297
FOUILLET A, REY G, LAURENT F, PAVILLON G, BELLEC S, GUIHENNEUC-
JOUYAUX C, CLAVEL J, JOUGLA E, and HÉMON D (2006) Excess mortality
related to the august 2003 heat wave in france. International Archives of Occupational
and Environmental Health 80 (1), pp. 16-24.
FRANKFORT-NACHMIAS C, and NACHMIAS D (1996) Research methods in social
sciences. Arnold, London.
FRENCH S, and TUROFF M (2007) Decision support systems. Commun. ACM 50 (3), pp.
39-40.
GALBRAITH JR (1974) Organization design: An information processing view. Interfaces 4
(3), pp. 28-36.
GALLIVAN M, and SRITE M (2005) Information technology and culture: Identifying
fragmentary and holistic perspectives of culture. Information and Organization 15 pp.
295-338.
GARCIA AC, STANDLEE AI, BECHKOFF J, and YAN CUI (2009) Ethnographic
approaches to the internet and computer-mediated communication. Journal of
Contemporary Ethnography 38 (1), pp. 52-84.
GARNETT JL, and KOUZMIN A (2007) Communicating throughout katrina: Competing and
complementary conceptual lenses on crisis communication. Public Administration
Review 67 pp. 171-188.
GASSON S (Year) Boundary-spanning knowledge-sharing in e-collaboration. In System
Sciences, 2005. HICSS '05. Proceedings of the 38th Annual Hawaii International
Conference onEds).
GAUTHEREAU V, and HOLLNAGEL E (2005) Planning, control, and apdaptation: A case
study. European Management Journal 23 (1), pp. 118.
GAVER WW "Technology affordances," in: Proceedings of the SIGCHI conference on
Human factors in computing systems: Reaching through technology, ACM, New
Orleans, Louisiana, United States, 1991.
GEORGE JM (1997) Organizational spontaneity in context. Human performance 10 (2), pp.
153.
GEPHART RP, JR. (1984) Making sense of organizationally based environmental disasters.
Journal of Management 10 (2), pp. 205-225.
GIOIA DA, and POOLE PP (1984) Scripts in organizational behavior. The Academy of
Management review 9 (3), pp. 449-459.
GLASER BG, and STRAUSS AL (1973) The discovery of grounded theory: Strategies for
qualitative research. Aldine Transaction, Chicago.271 p.
GOFFMAN E (1967) Interaction rituals. Garden City, New York.
GOFFMAN E (1970) Strategic interaction. University of Pennsylvania Press,
GOFFMAN E (1974) Frane analysis: An essay on the organization of experience. Harper &
Row,
GOLDEN BR (1992) The past is the past- or is it? The use of retropsective accounts as
indivators of past strategy. Academy of management 13 pp. 145-158.

  298
GOMBAULT A (2005) La méthode des cas. In: Management des ressources humaines:
Méthodes de recherche en sciences humaines et sociales, (ROUSSEL P and
WACHEUX E, Eds.), pp. 30-64, De Boeck université
GOODHUE DL, and THOMPSON RL (1995) Task-technology fit and individual
performance. MIS Quarterly 19 (2), pp. 213-236.
GOT C (2008) Les revers de la santé publique. Sève - Presses de Sciences Po 3 (20), pp. 23-
38.
GRANOVETTER MS (1973) The strength of weak ties. The American Journal of Sociology
78 (6), pp. 1360-1380.
GRAWITZ M (2001) Méthodes des sciences sociales. Dalloz, Paris.
GRIFFITH TL, SAWYER JE, and NEALE MA (2003) Virtualness and knowledge in teams:
Managing the love triangle of organizations, individuals, and information technology.
Mis Quarterly 27 (2), pp. 265-287.
GUHA-SAPIR D, HARGITT D, and HOYOIS P "Thirty years of natural disasters 1974-
2003: The numbers," Presses Universitaires de Louvain, Louvain, 2004.
HAIG BD (1995) Grounded theory as scientific method. In: Philosophy of educationEds.),
University of Illinois Press
HAIG BD (2005) An abductive theory of scientific method. Psychological Method 10 (4), pp.
371-388.
HAIR JJF, ANDERSON RE, TATHAM RL, and C. BW (1984) Multivariate data analysis.
Prentice Hall, Upper Saddle River, New Jersey.730 p.
HALE J, HALE DP, and DULEK RE (2006) Decision processes during crisis response: An
exploratory investigation. Journal of managerial issues 18 (3), pp. 301-320.
HALL DT (1971) Organizational and individual response to external stress. Administrative
Science Quarterly 16 (3), pp. 533.
HAMBRICK DC (1980) The concept of business-level strategy in research. The Academy of
Management Review 5 (4), pp. 567-575.
HANSETH O, MONTEIRO E, and HATLING M (1996) Developing information
infrastructure: The tension between standardization and flexibility. Science,
Technology, & Human Values 21 (4), pp. 407-426.
HARE AP, BORGATTA EF, and BALES RF (1965) Small groups studies in social
interaction. Knopff, A. A., New York.706 p.
HARRIS H (2001) Content analysis of secondary data: A study of courage in managerial
decision making. Journal of Business Ethics 34 pp. 191-208.
HART P (1993) Symbols, rituals and power: The lost dimensions of crisis management.
Journal of contingencies and crisis management 1 (1), pp. 36.
HATCH MJ (1998) Jazz as a metaphor for organizing in the 21st century. Organization
Science 9 (5), pp. 556-557.
HATCH MJ (1999) Exploring the empty spaces of organizing: How improvisational jazz
helps redescribe organizational structure. Organization Studies 20 (1), pp. 75-100.
HEEKS R (2002a) Information systems and developing countries: Failure, success, and local
improvisations. The Information Society 18 (2), pp. 101-112.

  299
HEEKS R (2002b) Information systems and developing countries: Failure, success, and local
improvisations. The Information Society 18 (2), pp. 101–112.
HEEKS R (2006) Health information systems: Failure, success and improvisation.
International Journal of Medical Informatics 75 (2), pp. 125-137.
HÉMON D, JOUGLA E, CLAVEL J, LAURENT F, BELLEC S, and PAVILLON G (2003)
Surmortalité liée à la canicule d'août 2003 en france. Bulletin Epidémiologique
Hebdomadaire 45-46 (221-225),
HENDERSON JC, and VENKATRAMAN N (1993) Strategic alignment: Leveraging
information technology for transforming organizations. IBM Syst. J. 32 (1), pp. 4-16.
HERMANN CF (1963) Some consequences of crisis which limit the viability of
organizations. Administrative Science Quarterly 8 (1), pp. 61-82.
HILTZ SR, and TUROFF M (1985) Structuring computer-mediated communication systems
to avoid information overload. Commun. ACM 28 (7), pp. 680-689.
HMIELESKI KM, and CORBETT AC (2008) The contrasting interaction effects of
improvisational behavior with entrepreneurial self-efficacy on new venture
performance and entrepreneur work satisfaction. Journal of Business Venturing 23 (4),
pp. 482-496.
HOLDEN PJP (2005) Improvising in an emergency. New England Journal of Medicine, The
353 (6), pp. 541.
HOOKWAY C (2004) The principle of pragmatism: Pierce's formulation and examples.
Midwest Studies in Philosophy pp. 119-136.
HUNINK MGM, and GLASZIOU PP (2001) Decision making in health care and medecine:
Integrating evidence and values. Cambridge University Press, Cambdrige.388 p.
HUTCHINS E (1991) Organizing work by adaptation. Organization Science 2 (1), pp. 14-39.
ISAAC H, CAMPOY E, and KALIKA M (2007) Surcharge informationnelle, urgence et tic.
L'effet temporel des technologies de l'information. (french). Revue Management et
Avenir (13), pp. 149-168.
IVES B, and JUNGLAS I (2006a) Information systems at northrop grumman ship systems
sector: The hurricane katrina recovery. Communications of the Associations for
Information Systems 18 pp. 557_577.
IVES B, and JUNGLAS I (2006b) Information systems at northrop grumman ship systems
sector: The hurricane katrina recovery. Communications of the Associations for
Information Systems 18 pp. 557-577.
JAEGER PT, SHNEIDERMAN B, FLEISCHMANN KR, PREECE J, QU Y, and FEI WU P
(2007) Community response grids: E-government, social networks, and effective
emergency management. Telecommunications Policy 31 (10-11), pp. 592-604.
JAMES W (1897) The will to believe and other essays in popular philosophy. Longmans,
Green & co,
JAMES W (1975) Pragmatism. Harvard University Press, 316 p.
JEFFERSON T (2006a) Evaluating the role of information technology in crisis & emergency
management. VINE:The journal of information and knowledge management systems
36 (3), pp. 261-264.

  300
JEFFERSON T (2006b) Using the internet to communicate during a crisis. VINE:The journal
of information and knowledge management systems 36 (2), pp. 139-142.
JENNEX ME (2008) Emergency response systems: The utility y2k experience. Journal of
Information Technology Theory and Application 9 (3), pp. 85-102.
JIYAO C, and JING M (Year) The role of improvisation in entrepreneurial processes. In
Engineering Management Conference, 2005. Proceedings. 2005 IEEE
InternationalEds).
JOFFRE P, AURÉGAN P, CHÉDOTEL F, and TELLIER A (2006) Le management
stratégique par le projet. Economica, Paris.
JOLIBERT A, and JOURDAN P (2006) Marketing research. Méthodes de recherche et
d'études en marketing. Dunod, Paris.
KALIKA M, BOUKEF CHARKI N, and ISAAC H (2007) La théorie du millefeuille et
l'usage des tic dans l'entreprise. Revue française de gestion 172 pp. 117-129.
KAMOCHE K, CUNHA MP, and CUNHA JV (2003) Towards a theory of organizational
improvization: Looking beyond the jazz metaphor. The Journal of management
studies 40 (8), pp. 2023-2051.
KAMOCHE K, and CUNHA PM (2001) Minimal structures: From jazz improvisation to
product innovation. Organization Studies 22 (5), pp. 733-764.
KAPUCU N (2006) Interagency communication networks during emergencies: Boundary
spanners in multiagency coordination. American Review of Public Administration 36
(2), pp. 207-225.
KATZ D, and KAHN RL (1966) The social psychology of organizations. John Wiley & Sons,
KHAZANCHI D (2005) Information technology (it) appropriateness: The contingency theory
of "Fit" And it implementation in small and medium enterprises. Journal of Computer
Information Systems 45 (3), pp. 88-95.
KING AW, and RANFT AL (2001) Capturing knowledge and knowing through
improvisation: What managers can learn from the thoracic surgery board certification
process. Journal of Management 27 (3), pp. 255-277.
KIRKMAN BL, and MATHIEU JE (2005) The dimensions and antecedents of team
virtuality. Journal of Management 31 (5), pp. 700-718.
KIRMAYER LJ (1994) Improvisation and authority in illness meaning. Culture, Medicine
And Psychiatry 18 (2), pp. 183-214.
KLEIN HK, and MYERS MD (1999) A set of principles for conducting and evaluating
interpretive field studies in information systems. MIS Quarterly 23 (1), pp. 67-93.
KNOLL K, and JARVENPAA SL "Information technology alignment or "Fit" In highly
turbulent environments: The concept of flexibility," in: Proceedings of the 1994
computer personnel research conference on Reinventing IS : managing information
technology in changing organizations: managing information technology in changing
organizations, ACM, Alexandria, Virginia, United States, 1994, pp. 1-14.
KNOX H, O'DOHERTY D, VURDUBAKIS T, and WESTRUP C (2007) Rites of passage:
Organization as an excess of flows. Scandinavian Journal of Management 23 (3), pp.
265-284.

  301
KOCK N, GRAY P, HOVING R, KLEIN H, MYERS MD, and ROCKART J (2002) Is
research relevance revisited: Subtle accomplishment, unfulfilled promise, or serial
hypocrisy? Communications of the Association for Information Systems 8 pp. 330-
346.
KOENIG G (1993) Production de la connaissance et constitution des pratiques
organisationnelles. Revue de Gestion des Ressources Humaines 9 pp. 4-17.
KONSYNSKI B, and TIWANA A (2004) The improvisation-efficiency paradox in inter-firm
electronic networks: Governance and architecture considerations. Journal of
Information Technology 19 (4), pp. 234-243.
KOONTZ H (1980) The management theory jungle revisited. The Academy of Management
Review 5 (2), pp. 175-187.
KREPS GA, and BOSWORTH SL (1993) Disaster, organizing, and role enactment: A
structural approach. American Journal of Sociology 99 (2), pp. 428-463.
KRISTENSEN M, KYNG M, and PALEN L (Year) Participatory design in emergency
medical service: Designing for future practice. In CHI 2006
ProceedingsEds)Miontréal, Québec, Canada.
KUHN TS (1962) The structures of scientific revolutions. University of Chicago Press,
Chicago.
LAGADEC P (2004) Understanding the french 2003 heat wave experience: Beyond the heat,
a multi-layered challenge. Journal of contingencies and crisis management 12 (4), pp.
160.
LAGADEC P (2005) Retour d'expérience: Théorie et pratique. La canicule de l'été 2003.
Auscultation des rapports d'enquête. In: Retour sur les rapports d'enquête et
d'expertise suite à la canicule de l'été 2003Eds.), pp. 17-200, CNRS/MSH Alpes,
Grenoble.
LALANDE F, LEGRAIN S, VALLERON A-J, MEYNIEL D, and FOURCADE M "Mission
d'expertise et d'évaluation du système de santé pendant la canicule 2003," Ministère de
la santé, de la famille et des personnes handicapées, Paris.
LANDGREN J "Designing information technology for emergency response," in: Department
of Applied Information Technology, University of Goteborg, Goteborg, 2007.
LANG G, and BENBUNAN-FICH R (2010) The use of social media in disaster situations:
Framework and cases. International Journal of Information Systems for Crisis
Response and Management 2 (1), pp. 11-23.
LANZARA GF (1983) Ephemeral organizations in extreme environments: Emergence,
strategy, extinction. Journal of Management Studies 20 (1), pp. 71-95.
LANZARA GF (1999) Between transient constructs and persistent structures: Designing
systems in action. The Journal of Strategic Information Systems 8 (4), pp. 331-349.
LARSON RC, METZGER MD, and CAHN MF (2006) Responding to emergencies: Lessons
learned and the need for analysis. INTERFACES 36 (6), pp. 486-501.
LATOUR B (2004) Changer de société, refaire de la sociologie. La Découverte, 400 p.
LAU F, DOZE S, VINCENT D, WILSON D, NOSEWORTHY T, HAYWARD R, and
PENN A (1999) Patterns of improvisation for evidence-based practice in clinical
settings. Information, Technology & People 12 (3), pp. 287-303.

  302
LAWRENCE BS (2006) Organizational reference groups: A missing perspective on social
context. Organization Science 17 (1), pp. 80.
LE FAOU A-L (2003) Les systèmes de santé en questions - allemagne, france, royaume-uni,
etats unis et canada. Ellipses, 224 p.
LE GRAND-SÉBILLE C, and VÉGA A (2005) Pour une autre mémoire de la canicule.
Vuibert,
LEE AS (1999) Rigor and relevance in mis research: Beyond the approach of positivism
alone. MIS Quarterly 23 (1), pp. 29-33.
LEIDNER D, PAN G, and PAN S "The multi-faceted role of it in resource deployment during
crisis response: Lessons from sars and asian tsunami disasters," 2009, p. 45.
LEIFER R, and HUBER G (1977) Relations among perceived uncertainty, organizational
structure, and boundary spanning behavior. Administrative Science Quaterly 22 pp.
235-247.
LÉVI-STRAUSS C (1962) La pensée sauvage. Plon, Paris.382 p.
LEWIS MW (2000) Exploring paradox: Toward a more comprehensive guide. The Academy
of Management review 25 (4), pp. 760-776.
LEYBOURNE S (2006) Improvisation within the project management of change: Some
observations from uk financial services. Journal of Change Management 6 (4), pp.
365.
LIN A, and CONFORD T "Framing implementation management," in: Proceedings of the
twenty first international conference on Information systems, Association for
Information Systems, Brisbane, Queensland, Australia, 2000.
LIN Z, and CARLEY KM (1997) Organizational response: The cost performance tradeoff.
Management science 43 (2), pp. 217-234.
LIN Z, ZHAO X, ISMAIL KA, and CARLEY KM (2006) Organizational design and
restructuring in response to crises: Lessons from computational modeling and real-
world cases. Organization Science 17 (5), pp. 598-618.
LINCOLN JR, and JON M (1979) Work and friendship ties in organizations: A comparative
analysis of relation networks. Administrative Science Quarterly 24 (2), pp. 181-199.
LIU SB, PALEN L, SUTTON J, HUGHES AL, and VIEWEG S (Year) In search of the
bigger picture: The emergent role of on-line photo sharing in times of disaster. In
Proceedings of the 5th International ISCRAM Conference, (FIEDRICH F and VAN
DE WALLE B, Eds)Washington, DC, USA.
LOCKE K (2001) Grounded theory in management research. Sage Editions.,
LYYTINEN K (1999a) Empirical research in information systems: On the relevance of
practice in thinking of is research. MIS Quarterly 23 (1), pp. 25-27.
LYYTINEN K (1999b) Empirical research on information systems: On the relevance of
practice in thinking of is research. MIS Quarterly 23 (1), pp. 25-28.
MACREDIE RD, and SANDOM C (1999) It-enabled change: Evaluating an improvisational
perspective. European Journal of Information Systems 8 (4), pp. 247-259.

  303
MAJCHRZAK A, JARVENPAA SL, and HOLLINGSHEAD AB (2007) Coordinating
expertise among emergent groups responding to disasters. Organization Science 18
(1), pp. 147-161.
MALIZIA A, ASTORGA-PALIZA F, ONORATI T, DIAZ P, and AEDO I (Year)
Emergency alerts for all: An ontology based approach to improve accessibility in
emergency alerting systems. In ISCRAM, (FRIEDRICH F and VAN DE WALLE B,
Eds)Washington, DC.
MANISERA M, DUSSELDORP E, and VAN DER KOOIJ AJ "Identifying the component
structure of job statisfaction by categorical principal component analysis," 2006, pp.
1-43.
MARCH JG, and SIMON HA (1958) Organizations. Wiley, New York.
MARINCIONI F (2007) Information technologies and the sharing of disaster knowledge: The
critical role of professional culture. Disasters 31 (4), pp. 459-476.
MARKUS ML, MAJCHRZAK A, and GASSER L (2002) A design theory for systems that
support emergent knowledge processes. MIS Quarterly 26 (3), pp. 180-212.
MARKUS ML, and ROBEY D (1988) Information technology and organizational change:
Causal structure in theory and research. Management Science 34 (5), pp. 583-598.
MARROU H-I (1954) De la connaissance historique. Editions du seuil, 318 p.
MASINO G, and ZAMARIAN M (2003) Information technology artefacts as structuring
devices in organizations: Design, appropriation and use issues. Interacting with
Computers 15 (5), pp. 693-707.
MASON J (2002) Qualitative researching. Sage,
MASON RO, MCKENNEY JL, and COPELAND DG (1997) Developing an historical
tradition in mis research. MIS Quarterly 21 (3), pp. 257-278.
MASSEY AP, and MONTOYA-WEISS MM (2006) Unraveling the temporal fabric of
knowledge conversion: A model of media selection and use. Mis Quarterly 30 (1), pp.
99-114.
MASSEY AP, MONTOYA-WEISS MM, and YU-TING H (2003) Because time matters:
Temporal coordination in global virtual project teams. Journal of management
information systems 19 (4), pp. 129-155.
MATHEW D (2005) Information technology and public health management of disasters--a
model for south asian countries. Prehospital and disaster medicine 20 (1), pp. 54-60.
MATHEWS KM, WHITE MC, and LONG RG (1999) The problem of prediction and control
in theoretical diversity and the promise of the complexity sciences. Journal of
Management Inquiry 8 (1), pp. 17-31.
MATHIESON K, and KEIL M (1998) Beyond the interface: Ease of use and task/technology
fit. Information & Management 34 (4), pp. 221-230.
MAULE AJ, and EDLAND AC (1997) The effects of time pressure on human judgment and
decision making. In: Decision making: Cognitive models and explanations,
(RANYARD R, CROZIER WR and SVENSON O, Eds.), Routledge
MCGANN S "Coping with the unplanned the dynamics of improvisation in information
systems evolution within and across firm boundaries," in: Department of Information
Systems, Weatherhead School of Management, 2005.

  304
MCGANN ST, and LYYTINEN K (Year) The improvisation effect: A case study of user
improvisation and its effects on information system evolution - paper 209. In
ICISEds)Paris.
MCGRATH JE (1991) Time, interaction, and performance (tip): A theory of groups. Small
group research 22 (2), pp. 147-174.
MCKINNEY EH (2008) Supporting pre-existing teams in crisis with it: A preliminary
organizational-team collaboration framework. Journal of Information Technology
Theory and Application 9 (3), pp. 39-59.
MCKNIGHT B, and BONTIS N (2002) E-improvisation: Collaborative groupware
technology expands the reach and effectiveness of organizational improvisation.
Knowledge and process management 9 (4), pp. 219-227.
MCMANUS ST "Organisational resilience in new zealand," University of Canterbury, 2008.
MEAD GH (1934) Mind, self, and society: From the standpoint of a social behaviorist.
University of Chiacgo Press,
MENDONCA D (2007) Decision support for improvisation in response to extreme events:
Learning from the response to the 2001 world trade center attack. Decision Support
Systems 43 (3), pp. 952-967.
MENDONÇA D (2007) Decision support for improvisation in response to extreme events:
Learning from the response to the 2001 world trade center attack. Decision Support
Systems 43 (3), pp. 952-967.
MENDONÇA D, BEROGGI GEG, VAN GENT D, and WALLACE WA (2006) Designing
gaming simulations for the assessment of group decision support systems in
emergency response. Safety Science 44 (6), pp. 523-535.
MENDONÇA D, BEROGGI GEG, and WALLACE WA (Year) Evaluating support for
improvisation in simulated emergency scenarios. In Hawaii International Conference
on System Sciences (HICSS)Eds).
MENDONÇA D, and FIEDRICH F (Year) Design for improvisation in computer-based
emergency response systems. In ISCRAMEds)Brussels.
MENDONÇA D, JEFFERSON T, and HARRALD J (2007) Collaborative adhocracies and
mix-and-match technologies in emergency management - using the emergent
interoperability approach to address unanticipated contingencies during emergency
response. Communications of the ACM 50 (3), pp. 45-49.
MENDONÇA D, LEE II. EA, and WALLACE WA (2004a) Impact of the 2001 world trade
center attack on critical interdependent infrastructures.
MENDONÇA D, and WALLACE WA (2004) Studying organizationally-situated
improvisation in response to extreme events. International Journal of Mass
Emergencies and Disasters 22 pp. 5-29.
MENDONÇA D, WEBB EJ, and BUTTS C (2010) L'improvisation dans les interventions
d'urgence: Les relations entre cognition, comportement et interactions sociales. Tracés
18 (1), pp. 69-86.
MENDONÇA DJ, and WALLACE WA (2007) A cognitive model of improvisation in
emergency management. Ieee Transactions on Systems Man and Cybernetics Part a-
Systems and Humans 37 (4), pp. 547-561.

  305
MENDONÇA S, CUNHA MP, KAIVO-OJA J, and RUFF F (2004b) Wild cards, weak
signals and organisational improvisation. Futures 36 (2), pp. 201-232.
MEYER A (1998) Organizing for improvisation: The backstage story of the vancouver jazz
concert and symposium. Organization Science 9 (5), pp. 569-576.
MEYER A, FROST PJ, and WEICK KE (1998) The organization science jazz festival:
Improvisation as a metaphor for organizing - overture. Organization Science 9 (5), pp.
540-542.
MILBURN TW, SCHULER RS, and WATMAN KH (1983a) Organizational crisis. Part i:
Definition and conceptualization. Human Relations 36 (12), pp. 1141-1160.
MILBURN TW, SCHULER RS, and WATMAN KH (1983b) Organizational crisis. Part ii:
Strategies and responses. Human Relations 36 (12), pp. 1161-1171.
MILES MB, and HUBERMAN AM (1994) Qualitative data analysis: An expanded
sourcebook. Sage Publications.,
MILET M (2005) Cadres de perception et luttes d'imputation dans la gestion de crise:
L'exemple de "La canicule" D'août 2003. Revue française de science politique 55 (4),
pp. 573-605.
MINER AS, BASSOFF P, and MOORMAN C (2001) Organizational improvisation and
learning: A field study. Administrative Science Quarterly 46 (2), pp. 304-337.
MINTZBERG H (1979) The structuring of organizations. Hardback, 512 p.
MINTZBERG H (1996) Structure et dynamique des organisations. Broché,
MINTZBERG H, and MCHUGH A (1985) Strategy formation in an adhocracy.
Administrative Science Quarterly 30 (2), pp. 160-197.
MISHRA A (1996) Organizational responses to crisis: The centrality of trust. In: Trust in
organizations: Frontiers of theory and research, (KRAMER RM and TYLER TR,
Eds.), pp. 261-261,
MITCHELL TR, and THOMPSON L (1994) A theory of temporal adjustments of the
evaluation of events:Rosy prospection & rosy prospection. In: Advances in
managerial cognition and information processing, (STUBBART C, MEINDL JR and
PORAC JF, Eds.), pp. 85-114, JAI Press, Greenwich.
MITROFF II, PAUCHANT T, FINNEY M, and PEARSON C (1989) Do (some)
organizations cause their own crises? The cultural profiles of crisis-prone vs. Crisis-
prepared organizations. Organization Environment 3 (4), pp. 269-283.
MITROFF II, PEARSON CM, and HARRINGTON LK (1996) The essential guide to
managing corporate crisis: A step-by-step handbook for surviving major catastrophes.
Oxford University Press,
MONTUORI A (2003) The complexity of improvisation and the improvisation of
complexity: Social science, art and creativity. Human Relations 56 (2), pp. 237-255.
MOORMAN C, and MINER AS (1998a) Organizational improvisation and organizational
memory. Academy of Management Review 23 (4), pp. 698-723.
MOORMAN C, and MINER AS (1998b) The convergence of planning and execution:
Improvisation in new product development. Journal of Marketing 62 (3), pp. 1-20.

  306
MORGAN J, and KRONE KJ (2001) Bending the rules of" Professional" Display: Emotional
improvisation in caregiver performances. Journal of applied communications research
29 (4), pp. 317-340.
MORIN EM (1976) Pour une crisiologie. Communications 25 pp. 149-163.
MORIN EM (1993) Enantiodromia and crisis management: A jungian perspective.
Organization & environment 7 (2), pp. 91-114.
MORK L (2002) Technology tools for crisis response. Risk management 49 (10), pp. 44.
MULLEN J, VLADI N, and MILLS AJ (2006) Making sense of the walkerton crisis. Culture
& Organization 12 (3), pp. 207-220.
NEWCOMB TM (1953) An approach to the study of communicative acts. Psychological
Review 60 (6), pp. 393-404.
NEWCOMB TM, TURNER RH, and CONVERSE PE (1965) Social psychology: The study
of human interaction. Holt, Rinehart, Winston, Inc., New York.
NEWCOMB TM, TURNER RH, and CONVERSE PE (1970) Manuel de psychologie
sociale: L'interaction entre individus. Presses Universitaires de France, Paris.
NTUEN C, BALOGUN O, BOYLE E, and TURNER A (2006) Supporting command and
control training functions in the emergency management domain using cognitive
systems engineering. Ergonomics 49 (12/13), pp. 1415-1436.
NUNAMAKER JR JF, WEBER ES, and CHEN M (1989) Organizational crisis management
systems: Planning for intelligent action. Journal of management information systems 5
(4), pp. 7-32.
NYSTRÖM PC, and STARBUCK WH (1984) To avoid organizational crisis: Unlearn.
Organizational dynamics 12 (4), pp. 53-65.
ORLIKOWSKI WJ (1992) The duality of technology: Rethinking the concept of technology
in organizations. Organization Science 3 (3), pp. 31.
ORLIKOWSKI WJ (1996) Improvising organizational transformation over time: A situated
change perspective. Information Systems Research 7 (1), pp. 63-92.
ORLIKOWSKI WJ (2000) Using technology constituting structure : Practical lens for
studying technology organizations. Organization Science 11 (4), pp. 25.
ORLIKOWSKI WJ (2007) Sociomaterial practices: Exploring technology at work.
Organization Studies 28 (9), pp. 1435-1448.
ORLIKOWSKI WJ, and HOFMAN JD (1997) An improvisational model for change
management: The case of groupware technologies. In: Inventing organizations of the
21st century: Producing knowledge through collaboration, (MALONE TW,
LAUBACHER R, SCOTT MORTON MS and MANAGEMENT SSO, Eds.), pp. 265-
282, MIT Press
ORLIKOWSKI WJ, and IACONO CS (2001) Research commentary: Desperately seeking the
"It" In it research - a call to theorizing the it artifact. Information Systems Research 12
(2), pp. 121-134.
ORLIKOWSKI WJ, and ROBEY D (1991) Information technology and the structuring in
organizations. Information Systems Research 2 (2), pp. 27.

  307
ORLIKOWSKI WJ, and YATES J (1994) Genre repertoire: The structuring of
communicative practices in organizations. Administrative Science Quarterly 39 (4),
pp. 541-574.
ORTON JD (1997) From inductive to iterative grounded theory: Zipping the gap between
process theory and process data. Scandinavian Journal of Management 13 (4), pp.
419-438.
OSBORN CS (1998) Systems for sustainable organizations: Emergent strategies, interactive
controls and semi-formal information. Journal of Management Studies 35 (4), pp.
481-509.
OUTHWAITE W (2003) Realism and social science. In: Critical realism: Essential readings,
(ARCHER M, BASHKAR R, COLLIER A, LAWSON T and NORRIE A, Eds.), pp.
282-296, Routledge
PALEN L, HILTZ SR, and LIU SB (2007) Online forums supporting grassroots participation
in emergency preparedness and response. Commun. ACM 50 (3), pp. 54-58.
PALEN L, and LIU SB (Year) Citizen communications in crisis: Anticipating a future of ict-
supported public participation. In CHI 2007 - Emergency ActionEds)San Jose, CA,
USA.
PALEN L, and LIU SB "Citizen communications in crisis: Anticipating a future of ict-
supported public participation," in: Proceedings of the SIGCHI conference on Human
factors in computing systems, ACM, San Jose, California, USA, 2007b.
PANTELI N, and FINEMAN S (2005) The sound of silence: The case of virtual team
organising. Behaviour & Information Technology 24 (5), pp. 347-352.
PAUCHANT TC, and DOUVILLE R (1993) Recent research in crisis management: A study
of 24 authors' publications from 1986 to 1991. Organization & environment 7 (1), pp.
43-66.
PAUCHANT TC, and MITROFF I (1992) Transforming the crisis-prone organization:
Preventing individual, organizational, and environmental tragedies. Jossey-Bass.,
PAUCHANT TC, and MITROFF II (1988) Crisis prone versus crisis avoiding organizations
is your company's culture its own worst enemy in creating crises? Organization
Environment 2 (1), pp. 53-63.
PAVA ML (2002) The path of moral growth. Journal of Business Ethics 38 (1/2), pp. 43-54.
PAVLOU PA, BHARADWAJ A, EL SAWY O, GUPTA A, and TALLON P (Year) Linking
information technology and dynamic capabilities: The elusive dancing partners?.
Paper 85. In ICIS Eds).
PAWLOWSKI SD, WATSON EF, and NENOV IN "Enterprise systems, minimal structures
of knowledge and organizational improvisation," in: Session F-4, 2004.
PEARSON CM, and CLAIRE JA (1998) Reframing crisis management. The Academy of
Management review 23 (1), pp. 59-76.
PEARSON CM, MISRA SK, CLAIRE JA, and MITROFF I (1997) Managing the
unthinkable. Organizational dynamics 26 (2), pp. 51-64.
PELLOUX P (2004) L'urgentiste. Fayard, 300 p.
PENTLAND BT, and RUETER HH (1994) Organizational routines as grammars of action.
Administrative Science Quarterly 39 (3), pp. 484-510.

  308
PERROW C (1984) Normal accidents.
PERROW C (1999) Normal accidents: Living with high-risk technologies. Princeton
University Press,
PERROW C (2006) Disasters ever more? Reducing u.S vulnerabilities. In: Handbook of
disaster research, (RODRIGUEZ H, QUARANTELLI EL and DYNES R, Eds.),
Springer, New York.
PERRY DC, TAYLOR M, and DOERFEL ML (2003) Internet-based communication in crisis
management. Management communication quarterly 17 (2), pp. 206-231.
PERRY LT (1991a) Strategic improvising: How to formulate and implement competitive
strategies in concert. Organizational Dynamics 19 (4), pp. 51.
PERRY RW "Emergency management: Principles and practice for loccal government,
international city management association," in: International City Management
Association, DRABEK T and HOETMER G (eds.), Washington, 1991b, pp. 201-224.
PERRY RW, and QUARANTELLI EL (2005) What is a disaster?: New answers to old
questions. Xlibris Corporation,
PESQUEUX Y "Mais de quoi peut-il donc être question quand on parle de "Pratiques"?," in:
Billets d'humeur, AGRH, 2009.
PETTERSSON M, RANDALL D, and HELGESON B (2004) Ambiguities, awareness and
economy: A study of emergency service work. Computer Supported Cooperative
Work (CSCW) 13 (2), pp. 125-154.
PFEFFER J (1993) Barriers to the advance of organizational science: Paradigm development
as a dependent variable. The Academy of Management Review 18 (4), pp. 599-620.
PIAGET J (1977) Etudes sociologiques. Droz, 361 p.
PIERCE C (2000) What pragmatism is (1905). In: Pragmatism and classical american
philosophy - essential readings and interpretive essays, (STUHR JJ, Eds.), Oxford
University Press, New York.
PINNINGTON A, MORRIS T, and PINNINGTON C (2003) The relational structure of
improvisation-a case illustration from corporate video production. International
studies of management & organization 33 (1), pp. 10.
PISANELLI DM, GANGEMI A, BATTAGLIA M, and CATENACCI C (2004) Coping with
medical polysemy in the semantic web: The role of ontologies. Medinfo pp. 416-419.
POPPER KR (1972) Objective knowledge. Oxford University Press, Oxford.
PORTER ME (1991) Towards a dynamic theory of strategy. Strategic Management Journal
12 (S2), pp. 95-117.
POUTOUT G (2005) Réseaux de santé: Créer du lien pour donner du sens. Sociologies
pratiques 11 (2), pp. 33-54.
PRESCOTT MB, and CONGER SA (1995) Information technology innovations: A
classification by it locus of impact and research approach. SIGMIS Database 26 (2-3),
pp. 20-41.
PRESTON A (1987) Improvising order. In: Organization analysis and development: A social
construction of organizational, (MANGHAM IL, Eds.), Wiley, New York, NY.

  309
PROBST G, and RAISCH S (2005) Organizational crisis: The logic of failure. The Academy
of Management executive 19 (1), pp. 90-105.
PROST A (1996) Douze leçons sur l'histoire. Editions du Seuil, 330 p.
QUARANTELLI EL (1988) Disaster crisis management: A summary of research findings.
Journal of Management Studies 25 (4), pp. 373-385.
QUARANTELLI EL (1997) Problematic aspects of the information/communication
revolution for disaster planning research: Ten non-technical issues and questions.
Disaster prevention and management 6 (2), pp. 94-106.
QUARANTELLI EL "Disaster planning, emergency management and civil protection: The
historical development of organized efforts to plan for and to respond to disasters."
QUARANTELLI EL "Catastrophes are different from disasters: Some implications for crisis
planning and managing drawn from katrina," in: Understanding Katrina, Perspectives
from the Social Sciences, 2006.
QUARANTELLI EL "Problematic aspects of the computer based information/communication
revolution with respect to disaster planning and crisis managing," in: Preliminary
Work #358, University of Delaware Disaster Research Center, 2007.
QUARANTELLI EL, and DYNES RR (1977) Response to social crisis and disaster. Annual
review of sociology 3 (1), pp. 23-49.
RAFAELI S, and SUDWEEKS F (1997) Networked interactivity. Journal of Computer-
Mediated Communication 2 (4), pp. 1-6.
RAMESH B (1998) Factors influencing requirements traceability practice. Commun. ACM 41
(12), pp. 37-44.
RAMESH B, TIWANA A, and MOHAN K (2002) Supporting information product and
service families with traceability. In: Software product-family engineeringEds.), pp.
281-302,
REN Y, KIESLER S, and FUSSELL SR (2008) Multiple group coordination in complex and
dynamic task environments: Interruptions, coping mechanisms, and technology
recommendations. Journal of management information systems 25 (1), pp. 26.
RERUP C (2001) "Houston, we have a problem": Anticipation and improvisation as sources
of organizational resilience. Comportamento organizacional e gestao 7 (1), pp. 27-44.
RICE RE (1992) Task analyzability, use of new media, and effectiveness: A multi-site
exploration of media richness. Organization Science 3 (4), pp. 475-500.
RICHARD J (2008) Canicule. Une journée pour l'ignorance. Editions Bénévent, Nice.
RICHARDSON B (1994) Crisis management and management strategy–time to “loop the
loop”? Management 3 (3), pp. 59.
RIESKAMP J, and HOFFRAGE U (2008) Inferences under time pressure: How opportunity
costs affect strategy selection. Acta Psychologica 127 pp. 258-276.
ROBELET M, SERRÉ M, and BOURGUEIL Y (2005) La coordination dans les réseaux de
santé: Entre logiques gestionnaires et dynamiques professionnelles. Revue Française
des affaires sociales 1 (1), pp. 233-260.
ROBERT B, and LAJTHA C (2002) A new approach to crisis management. Journal of
contingencies and crisis management 10 (4), pp. 181-189.

  310
ROBERTS KH (1990a) Managing high reliability organizations. California Management
Review 32 (4), pp. 101-113.
ROBERTS KH (1990b) Some characteristics of one type of high reliability organization.
Organization Science 1 (2), pp. 160-176.
ROBERTS KH, and BEA R (2001) Must accidents happen? Lessons from high-reliability
organizations. Academy of Management Executive 15 (3), pp. 70-78.
ROBEY D (1996) Research commentary: Diversity in information systems research: Threat,
promise, and responsibility. Information Systems Research 7 (4), pp. 400-408.
ROBEY D, and SAHAY S (1996) Transforming work through information technology: A
comparative case study of geographic information systems in county government.
Information Systems Research 7 (1), pp. 93-110.
ROMELAER P "Organisation : Panorama d'une méthode de diagnostic," Laboratoire Crepa -
Université Paris Dauphine, Paris, p. 43.
ROSENTHAL U, and KOUZMINE A (1993) Globalizing an agenda for contingencies and
crisis management: An editorial statement. Journal of contingencies and crisis
management 1 (1), pp. 12.
ROUX-DUFORT C "L'apprentissage organisationnel post-crise," Université Paris-Dauphine,
Paris, 1997, p. 753.
ROUX-DUFORT C (2007) Is crisis management (only) a management of exceptions?
Journal of contingencies and crisis management 15 (2), pp. 105-114.
ROUX-DUFORT C, and VIDAILLET B (2003a) The difficulties of improvising in a crisis
situation - a case study. International studies of management & organization 33 (1),
pp. 86-115.
ROUX-DUFORT C, and VIDAILLET B (2003b) The difficulties of improvising in a crisis
situation-a case study. International studies of management & organization 33 (1), pp.
86-115.
RUESCH J, and BATESON G (1951) Communication the social matrix of psychiatry.
Norton, New York.
RUSSO E, RAPOSO A, FERNANDO T, GATTASS M, and KARLSSON B (Year)
Configuring a collaborative virtual workspace for disaster management on oil&gas
offshore structures. In Joint International Conference on Computing and Decision
Making in Civil and Building EngineeringEds)Montréal, Canada.
RYLE G (1979) On thinking. Basil Blackwell, Oxford.
SALAGNAC J-L (2007) Lessons from the 2003 heat wave: A french perspective. Building
Research & Information 35 (4), pp. 450 - 457.
SANDELANDS LE, and SRIVATSAN V (1993) The problem of experience in the study of
organizations. Organization Studies (Walter de Gruyter GmbH & Co. KG.) 14 (1), pp.
1-22.
SCHOENBERG A (2005) Do crisis plans matter? A new perspective on leading during a
crisis. Public relations quarterly 50 (1), pp. 2.
SCHULTZE U, and ORLIKOWSKI WJ (2001) Metaphors of virtuality: Shaping an emergent
reality. Information and Organization 11 (1), pp. 45-77.

  311
SEEGER MW, ULMER RR, NOVAK JM, and SELLNOW T (2005) Post-crisis discourse
and organizational change, failure and renewal. Journal of Organizational Change
Management 18 (1), pp. 78-95.
SEGEV A (Year) Adaptive ontology use for crisis knowledge representation. In The 5th
International ISCRAM Conference, (FRIEDRICH F and VAN DE WALLE B,
Eds)Washington, DC.
SHALUF IM, AHMADUM F, and SAID A (2003) A review of disaster and crisis. Disaster
prevention and management 12 (1), pp. 24-32.
SHARKANSKY I, and ZALMANOVITCH Y (2000) Improvisation in public administration
and policy making in israel. Public Administration Review 60 (4), pp. 321-329.
SHAW P, and STACEY RD (2006) Experiencing risk, spontaneity and improvisation in
organizational change: Working live. Taylor & Francis,
SHRIVASTAVA P (1987) Bhopal: Anatomy of a crisis. Paul Chapman Publishing, London.
SHRIVASTAVA P (1988) Industrial crisis management: Learning from organizational
failures. The Journal of management studies 25 (4), pp. 283-284.
SHRIVASTAVA P (1993) Crisis theory/practice: Towards a sustainable future. Organization
& environment 7 (1), pp. 23.
SHRIVASTAVA P (1994) The evolution of crisis research. Journal of contingencies and
crisis management 2 (1), pp. 10.
SHRIVASTAVA P, MITROFF I, MILLER D, and MIGLIANI A (1988a) Understanding
industrial crisis. The Journal of management studies 25 (4), pp. 285.
SHRIVASTAVA P, MITROFF II, MILLER D, and MIGLIANI A (1988b) Understanding
industrial crises [1]. The Journal of management studies 25 (4), pp. 19.
SIMON HA (1996) The sciences of the artificial. The MIT Press,
SMART C, and VERTINSKY I (1977) Designs for crisis decision units. Administrative
Science Quarterly 22 (4), pp. 640-657.
SMART C, and VERTINSKY I (1984) Strategy and the environment: A study of corporate
responses to crises. Strategic management journal 5 (3), pp. 199-213.
SMITH D (1990) Beyond contingency planning: Towards a model of crisis management.
Organization & environment 4 (4), pp. 263.
SOUSA CAA, and HENDRIKS PHJ (2006) The diving bell and the butterfly: The need for
grounded theory in developing a knowledge-based view of organizations.
Organizational Research Methods 9 (3), pp. 315-338.
STARBUCK WH, GREVE A, and HEDBERG B (1978) Responding to crisis. Journal of
business administration 9 (2), pp. 112-137.
STAW BM, SANDELANDS LE, and DUTTON JE (1981) Threat-rigidity effects in
organizational behavior: A multilevel analysis. Administrative Science Quarterly 26
(4), pp. 501-524.
STEAD E, and SMALLMAN C (1999) Understanding business failure: Learning and un-
learning from industrial crise. Journal of contingencies and crisis management 7 pp.
1-18.

  312
STRAUSS AL, and CORBIN J (2008) Basics of qualitative research: Techniques and
procedures for developing grounded theory. Sage publications, 379 p.
SUCHMAN L (1993) Response to vera and simon's situated action: A symbolic
interpretation. Cognitive Science 17 (1), pp. 71-75.
SUCHMAN L (2007) Human- machine reconfigurations: Plans and situated action.
Cambridge University Press,
SUCHMAN L, BLOMBERG J, ORR JE, and TRIGG R (1999) Reconstructing technologies
as social practice. American Behavioral Scientist 43 (3), pp. 392-408.
TEECE DJ, PISANO G, and SHUEN A (1997) Dynamic capabilities and strategic
management. Strategic management journal 18 (7), pp. 509-533.
THIÉTART R-A (2007) Méthodes de recherche en management Dunod, Paris.586 p.
THOMAS MA, ANDOH-BAIDOO FK, REDMOND R, and YOON V (2009) Moving
beyond traditional emergency response notification with voice xml. Journal of
Information Technology Theory and Application 10 (1), pp. 28-40.
TORRY WI (1979) Anthropological studies in hazardous environments: Past trends and new
horizons. Current Anthropology 20 (3), pp. 517.
TRAN S (2010) Quand les tic réussissent trop bien dans les organisations: Le cas du courrier
électronique chez les managers. Management & Avenir 34 (4), pp. 200-215.
TSAI W, and WU C-H "Knowledge combination: A co-citation analysis," in: Academy of
Management Journal, 2010, pp. 441-450.
TURNER BA (1976) The organizational and interorganizational development of disasters.
Administrative Science Quarterly 21 (3), pp. 378-397.
TURNER JH (1988) A theory of social interaction. Stanford University Press, Stanford,
California.
TUROFF M, CHUMER M, VAN DE WALLE B, and YAO X (2004) The design of a
dynamic emergency response management information system. Journal of
Information Technology Theory and Application 5 (4), pp. 1-35.
TUSHMAN ML, and SCANLAN TJ (1981) Boundary spanning individuals: Their role in
information transfer and their antecedents. Academy of Management Journal 24 (2),
pp. 289-305.
TYLERN JR, and TANG JC "When can i expect an email response? A study of rhythms in
email usage," in: Proceedings of the eighth conference on European Conference on
Computer Supported Cooperative Work, Kluwer Academic Publishers, Helsinki,
Finland, 2003.
VAAST E, and WALSHAM G (2005) Representations and actions: The transformation of
work practices with it use. Information and Organization 15 (1), pp. 65-89.
VAN DE VEN AH, DELBECQ AL, and KOENIG RJ (1976) Determinants of coordination
modes within organizations. American Sociological Review 41 (April), pp. 322-338.
VAN DE WALLE B, and RUTKOWSKI A-F (2006) A fuzzy decision support system for it
service continuity threat assessment. Decision Support Systems 42 (3), pp. 1931-1943.

  313
VAN DE WALLE B, and TUROFF M (2007) Special issue: Emergency reponse information
systems: Emerging trends and technologies. Communications of the ACM 50 (3), pp.
3.
VENDELØ MT (2009) Improvisation and learning in organizations - an opportunity for
future empirical research. Management Learning 40 (4), pp. 449-456.
VERA D, and CROSSAN M (2004) Theatrical improvisation: Lessons for organizations.
Organization Studies 25 (5), pp. 727-749.
VERA D, and CROSSAN M (2005) Improvisation and innovative performance in teams.
Organization Science 16 (3), pp. 203-224.
VERA D, and CROSSAN M (Year) Reconciling learning paradoxes through improvisation.
In OLKCEds).
VIEWEG S, PALEN L, LIU S, HUGHES AL, and SUTTON J (Year) Collective intelligence
in disaster: Examination of the phenomenin in the aftermath of the 2007 virginia
shooting. In the 5th International ISCRAM ConferenceEds)Washington, DC.
WAGNER EL, NEWELL S, and PICCOLI G (2010) Understanding project survival in an es
environment: A sociomaterial practice perspective. Journal of the Association for
Information Systems 11 (5), pp. 276-297.
WANG CL, and AHMED PK (2007) Dynamic capabilities: A review and research agenda.
International Journal of Management Reviews 9 (1), pp. 31-51.
WARKENTIN ME, SAYEED L, and HIGHTOWER R (1997) Virtual teams versus face-to-
face teams: An exploratory study of a web-based conference system*. Decision
Sciences 28 (4), pp. 975-996.
WATZLAWICK P, BEAVIN JH, and JACKSON DD (1967) Pragmatics of human
communication: A study of interaction patterns, pathologies, and paradoxes. Norton
& Company, New York.
WAUGH WL, and STREIB G (2006) Collaboration and leadership for effective emergency
management. Public Administration Review 66 pp. 131-140.
WEBB EJ (1996) Trust and crisis. In: Trust in organizations: Frontiers of theory and
research, (KRAMER RM and TYLER TR, Eds.), pp. 288-288,
WEBB EJ (2004) Role improvising in crisis response. International Journal of Emergency
Management 2 (1), pp. 47-61.
WEBB GR, and F-R. C (2006) Planning to improvise: The importance of creativity and
flexibility in crisis response. International Journal of Emergency Management 3 (1),
pp. 7.
WEBSTER J "Group spontaneity," in: Proceedings of the 1992 ACM SIGCPR conference on
Computer personnel research, ACM, Cincinnati, Ohio, United States, 1992.
WEBSTER J, and WATSON RT (2002) Analyzing the past to prepare for the future: Writing
a literature review. MIS Quarterly 26 pp. 13-23.
WEENIG MWH, and MAARLEVELD M (2002) The impact of time constraint on
information search strategies in complex choice tasks. Journal of Economic
Psychology 23 (6), pp. 689-702.
WEGNER DM (1987) Transactive memory: A contemporary analysis of the group mind. In:
Theories of group behaviorEds.), pp. 185-185,

  314
WEICK E, and SUTCLIFFE KM (2001) Managing the unexpected: Assuring high
performance in an age of complexity. University of Michigan Business School, 200 p.
WEICK KE (1988) Enacted sensemaking in crisis situations. The Journal of management
studies 25 (4), pp. 305-317.
WEICK KE (1993) The collapse of sensemaking in organizations - the mann gulch disaster.
Administrative Science Quarterly 38 (4), pp. 628-652.
WEICK KE (1998) Improvisation as a mindset for organizational analysis. Organization
Science 9 (5), pp. 543-555.
WEICK KE (2001a) Making sense of the organization. Blackwell Publisher,
WEICK KE (2001b) Making sense of the organization. Wiley-Blackwell, 483 p.
WEICK KE (2001c) Making sense of the organization. Blackwell Publishing, Malden.483 p.
WEICK KE (2006) Faith, evidence, and action: Better guesses in an unknowable world.
Organization Studies 27 (11), pp. 1723-1736.
WEICK KE, and ROBERTS KH (1993) Collective mind in organizations - heedful
interrelating on flight decks. Administrative Science Quarterly 38 (3), pp. 357-381.
WELLS K (1995) The strategy of grounded theory: Possibilities and problems. Social Work
Research 19 (1), pp. 33-37.
WENGER E (1999) Communities of practice: Learning, meaning, and identity. Cambridge
University Press,, Cambridge.318 p.
WHITTAKER S, and SIDNER C "Email overload: Exploring personal information
management of email," in: Proceedings of the SIGCHI conference on Human factors
in computing systems: common ground, ACM, Vancouver, British Columbia, Canada,
1996.
WINKIN Y (2001) Antropologie de la communication - de la théorie au terrain. Editions De
Boeck & Larvier S.A./Editions du Seuil,
YATES J, and ORLIKOWSKI W (2002) Genre systems: Structuring interaction through
communicative norms. Journal of Business Communication 39 (1), pp. 13-35.
YATES J, and ORLIKOWSKI WJ (1992) Genres of organizational communication: A
structurational approach to studying communication and media. The Academy of
Management Review 17 (2), pp. 299-326.
YATES J, ORLIKOWSKI WJ, and OKAMURA K (1999) Explicit and implicit structuring of
genres in electronic communication: Reinforcement and change of social interaction.
Organization Science 10 (1), pp. 83-103.
YIN RK (2003) Applications of case study research. Sage,
YUAN Y, and DETLOR B (2005) Intelligent mobile crisis response systems. Commun. ACM
48 (2), pp. 95-98.
YUAN-HSIANG L, JAN IC, KO PCI, YEN-YU C, JAU-MIN W, and GWO-JEN J (2004) A
wireless pda-based physiological monitoring system for patient transport. Information
Technology in Biomedicine, IEEE Transactions on 8 (4), pp. 439-447.
ZACK MH (2000) Jazz improvisation and organizing: Once more from the top. Organization
Science 11 (2), pp. 227-234.

  315
ZIGURS I, and BUCKLAND BK (1998) A theory of task/technology fit and group support
systems effectiveness. MIS Quarterly 22 (3), pp. 313-334.
ZIGURS I, and KHAZANCHI D (2008) From profiles to patterns: A new view of task-
technology fit. Information systems management 25 (1), pp. 8 - 13.
ZOLLO M, and WINTER SG (2002) Deliberate learning and the evolution of dynamic
capabilities. Organization Science 13 (3), pp. 339-351.
ZUCCHERMAGLIO C, and TALAMO A (2003) The development of a virtual community of
practices using electronic mail and communicative genres. Journal of Business and
Technical Communication 17 (3), pp. 259-284.

  316
Vu : le président

Vu : les suffragants

Vu et permis d’imprimer :
Le Vice-président du conseil chargé de la recherche de l’université Paris Dauphine.

  317